NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Intrapartum Care: Care of Healthy Women and Their Babies During Childbirth. London: National Institute for Health and Care Excellence (UK); 2014 Dec. (NICE Clinical Guidelines, No. 190.)

Cover of Intrapartum Care

Intrapartum Care: Care of Healthy Women and Their Babies During Childbirth.

Show details

3Place of birth

3.1. Introduction

Before 1945, the majority of births took place at home. The Cranbrook Report of 1959 stated that hospital maternity services were to provide for 70% of births, and in the 1960s hospitalisation of birth accelerated so that by 1970 nearly 90% of births occurred within hospitals.19,20 The Peel report in 1970 stated that facilities should be provided for all women to give birth in hospital, based largely on findings from the Reports of the Confidential Enquiry into Maternal Deaths, and this led rapidly to over 95% of women giving birth in a hospital setting.21 This provision of care was challenged and a number of initiatives culminated in the publication of the document Changing Childbirth in 1993 which recommended that women should have more choice in their place of birth, and that more choices should be available.22 The National Service Framework (NSF) for Children, Young People and Maternity Services in 2004 and Maternity Matters in 200725 both actively promoted midwife-led care for women, following appropriate assessment, and recommended that healthcare providers should develop midwife and home birth services to meet the needs of local populations.23,24 None of these initiatives were supported by strong evidence regarding safety of place of birth.

The configuration and choice of services are evolving, but more than 90% of births still take place in designated consultant wards (obstetric units) or combined consultant/GP wards.25 This figure is taken from the Maternity Hospital Episode Statistics but the categories used do not reflect current changes in practice. Also, local variation in the availability of different birth settings will affect women's options for choosing their preferred place of birth.

This section was prioritised for update following the publication of a large observational study conducted in England – Birthplace (2011) – which sought to answer the questions posed in a key research recommendation from the original Intrapartum care guideline. This study, plus a number of additional studies, have been incorporated into this update of the evidence, highlighting the continuing importance of women's choice in relation to this central component of intrapartum care.

3.2. Benefits and risks associated with each planned place of birth

3.2.1. Review question

What are the maternal and neonatal outcomes associated with planning birth in each of the following settings:

  • home (domiciliary)
  • freestanding midwifery unit
  • alongside midwifery unit
  • obstetric unit/hospital-based maternity unit.

For further details on the evidence review protocol, please see appendix E.

3.2.2. General points to note

The included studies report their outcomes in slightly different ways, so in some cases it was necessary to combine them in order to facilitate analysis. This occurs most commonly for the following outcomes:

  • Instrumental vaginal birth: where ventouse and forceps have been reported as two outcomes, these have been pooled because many studies do not report them separately.
  • Caesarean section: in some of the studies evaluating booked place of birth, there are a small proportion of elective caesarean sections as well as emergency caesarean sections.
  • Vaginal and perineal tears: due to the variation in how this is reported, these have been classified as any tears or third/fourth degree tears.

Full details of the individual outcomes reported in the studies can be found in the evidence tables (appendix I).

One study (Birthplace in England Collaborative Group, 2011) is included in all of the systematic reviews for this question. The following details should be noted about this study:

  • The above reference constitutes the main published paper for the study, but there are also further details noted in a more comprehensive report (Hollowell et al., 2011). This report was primarily used as a source of data for the outcome of maternal mortality (which was not reported elsewhere), for results of a subgroup analysis based on parity (multiparous and nulliparous) and for some details about reasons for transfer. Where it has been used as a source of data, this has been noted in the evidence tables.
  • For its primary outcome, Birthplace in England Collaborative Group, 2011, reports a post-hoc sub-group analysis, planned after data collection but before full analysis, for women without complicating conditions at the onset of labour, and this is reported in the evidence profiles in this document. For the remaining outcomes, the subgroup analysis was reported in the appendices of Hollowell et al. (2011) and the authors reported that the findings for women without complicating conditions at the start of care in labour were consistent with those for the whole study population.

3.2.3. Home compared with freestanding midwifery unit

3.2.3.1. Description of included studies

Two studies (reported in 3 publications) were included in this review (Birthplace in England Collaborative Group, 2011; Davis et al., 2011 and 2012).

One of the included studies was a prospective cohort study from England (Birthplace in England Collaborative Group, 2011). The second was a retrospective cohort study from New Zealand (Davis et al., 2011 and 2012).

Both studies compared planned birth at home with planned birth at a freestanding midwifery unit, and analysed data on an intention-to-treat basis, so that women were analysed by their planned place of birth even if they were transferred. Both studies evaluated intended place of birth at the onset of labour.

These studies were not pooled as they were observational data.

A summary of points to note about the study populations can be found in table 6 below, and further details about the selection of the study groups are reported in the evidence tables (appendix I).

Table 6. Summary of included studies for planned birth at home compared with planned birth in a freestanding midwifery unit.

Table 6

Summary of included studies for planned birth at home compared with planned birth in a freestanding midwifery unit.

3.2.3.2. Evidence profile

All risk ratios were calculated as standard using RevMan, but where the authors have reported adjusted measures of effect these have also been reported in the table. For measures of perinatal/neonatal mortality and morbidity, due to the low incidence absolute effects have been reported per 1,000,000.

Table 7. Summary GRADE profile for comparison of planned birth at home with planned birth in a freestanding midwifery unit for all women.

Table 7

Summary GRADE profile for comparison of planned birth at home with planned birth in a freestanding midwifery unit for all women.

3.2.3.3. Evidence statements

Only 2 studies (n=32,856) reported this comparison. The evidence around mode of birth was slightly inconsistent in terms of the magnitude of effect but the trends from both studies were that women planning birth at home had a lower rate of instrumental vaginal birth and a lower rate of caesarean section, and therefore a higher chance of a spontaneous vaginal birth, than women planning birth in a freestanding midwifery unit. Women planning birth at home also had lower rates of epidural use (n=28,050) but there was no difference in blood loss, either in terms of the risk of major postpartum haemorrhage (n=4,734) or the need for a blood transfusion (n=27,917). There was consistent evidence that women planning birth at home had lower rates of perineal trauma, either in the form of episiotomy (n=27,945), any perineal tears or third or fourth degree perineal tears (n=28,062).

In terms of neonatal outcomes, one study (n=27,752) found evidence of no difference in a composite adverse neonatal outcome between babies born to women planning birth at home and babies born to women planning birth in a freestanding midwifery unit midwifery unit. However, when sub-group analysis by parity was reported (n=9,646), the evidence from this study suggested that for nulliparous women there was a higher risk of a composite adverse neonatal outcome for babies whose mothers planned birth at home compared with those born to mothers who planned birth in a freestanding midwifery unit. There was no difference noted between groups for babies born to multiparous women. One study (n=27,799) reported neonatal encephalopathy and did not find a difference in risk for babies born to women planning birth at home and women planning birth in a freestanding midwifery unit, but this formed part of the composite outcome and the study was not powered to detect a difference in the individual components. Similarly, there was no evidence of a difference in stillbirth (n=32,820) and early neonatal death (n=32,721), but neither study was powered to detect a difference in these rare outcomes. There was no evidence of a difference in the risk of admission to a neonatal intensive care unit (NICU) (n=27,953) between the 2 groups of babies.

Within the 2 studies (n=32,856) overall the transfer rates were similar in both settings in each study. However, in 1 study only 1 in 10 women were transferred whereas in the other study about 1 in 5 women required transfer, and about three-quarters of those required transfer before birth. In that study transfers were much more common in nulliparous women in both birth settings. Just under 50% of transfers in multiparous women took place after birth. The evidence across all outcomes was of low and very low quality.

3.2.3.4. Evidence to recommendations

3.2.3.4.1. Relative value placed on the outcomes considered

The guideline development group members agreed that it was vital to consider both the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in women's decision-making process. For the baby, they felt that it was important to establish whether there was a difference in risk associated with planning birth at home compared with planning birth in a freestanding midwifery unit (given that they are both out-of-hospital settings) and therefore the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. Similarly, for the woman, the group wanted to ascertain whether there were differences in morbidity following planned birth at home compared with planned birth in a freestanding midwifery unit, given that there may be different facilities (and potentially more staff) available in the midwifery unit but that in the case of an emergency, a transfer by vehicle would be needed in either setting. The rates of intervention, such as caesarean section and instrumental vaginal birth, were also considered priorities, as they were felt to be important to women and also associated with morbidity, such as postpartum haemorrhage. The group did not feel that use of epidural anaesthesia was a particularly helpful outcome, as it is a matter of personal choice for the woman and would require a transfer from either setting.

The guideline development group members also felt that the rates of transfer were important to consider and would be an important consideration for women planning where to give birth.

3.2.3.4.2. Consideration of clinical benefits and harms

The guideline development group discussed the fact that the evidence around mode of birth varied between the two included studies. The Birthplace study suggested that the rates of spontaneous vaginal birth were higher, and the rates of instrumental vaginal birth and caesarean section were lower, after a planned home birth. In contrast, Davis et al. (2011 and 2012) did not find a significant difference between the two settings, although the direction of the effect was the same. The group noted that Davis et al. (2011 and 2012) had adjusted for differences in confounders between the two groups of women. Birthplace similarly adjusted for confounders and noted that the main difference in outcomes was found in nulliparous women.

The group noted that the evidence from the two studies for outcomes relating to blood loss was consistent; with Davis et al. (2011 and 2012) and Birthplace both reporting no significant difference in the rates of postpartum haemorrhage and blood transfusion respectively after planned home birth and planned birth in a freestanding midwifery unit. Similarly, the evidence for rates of any perineal tears, third or fourth degree perineal tears and episiotomy consistently showed that these outcomes were less common after planned birth at home.

The group noted that no statistically significant differences were identified for the neonatal outcomes reported but they did discuss the fact that neither of the studies was powered to detect differences in rare outcomes (stillbirth, neonatal death and neonatal encephalopathy). They noted that the Birthplace study was powered to detect a difference in its composite outcome, but that this had required pooling components with very different levels of severity, from fractured clavicle to mortality. While the group agreed that this was a limitation of the data, they also conceded that conducting a study with the power to detect differences in the very severe (and hence rare) outcomes would not be feasible. Also, fractured clavicle accounted for less than 3% of events and the serious perinatal outcomes – mortality, neonatal encephalopathy and meconium aspiration – constituted just under 90% of the primary outcome events. Given this evidence, and the fact that there was no significant difference in the rate of admission to NICU, the group noted that, overall, when parity was not taken in account, there did not appear to be a difference in neonatal outcomes between planned birth at home and planned birth in a freestanding midwifery unit. However, the group did note that parity was a large confounder in the study. This, and the fact that only 27.2% of the planned home birth group were nulliparous, means the overall result is likely to be misleading.

The guideline development group noted that in the Birthplace study, the overall rates of transfer from home and freestanding units were similar. Similarly, in both settings, nulliparous women were almost four times as likely to be transferred as multiparous women. Furthermore, nearly half of the transfers in multiparous women took place after birth. In Davis et al. (2011 and 2012), rates of transfer were lower in the planned birth in a midwifery unit group, and lower overall than those in the Birthplace study.

3.2.3.4.3. Consideration of health benefits and resource uses

The guideline development group members discussed the relative costs associated with planning birth at home and in a freestanding midwifery unit. They agreed that a home birth was likely to be cheaper due to the overhead costs associated with running a freestanding midwifery unit, and because the transfer rates (associated with the cost of an ambulance) were similar between the two settings. However, they also noted that in a freestanding midwifery unit the midwives can be supported by auxiliary staff in providing one-to-one care, but this is not possible in a planned home birth.

3.2.3.4.4. Quality of evidence

Only two studies reported this comparison and both were observational studies, but they were well conducted, fairly large, and published in 2011 and 2012. The Birthplace study was the larger of the two, was prospective and was conducted in England. However, there were some demographic differences between the two study groups which could have affected birth outcomes. The authors performed adjustments for these demographic differences for other comparisons reported in the study, but not for the comparison of home and freestanding midwifery units: as a result, the evidence was graded as very low quality. Davis et al. (2011 and 2012) was based in New Zealand, was retrospective and had a smaller sample size, but it did adjust for demographic differences between the study groups for most of its reported outcomes. The guideline development group noted that New Zealand has a higher proportion of both home and midwifery unit births than the UK and that this was a consideration, but they concluded overall that the evidence was applicable to women in England and Wales.

3.2.3.4.5. Other considerations

The guideline development group discussed the subgroup analysis by parity that was reported in the Birthplace study and noted some examples where outcomes were slightly different to the overall analysis. Whereas overall rates of spontaneous vaginal birth, caesarean section and third or fourth degree perineal trauma were lower for women who planned birth at home compared with those who planned birth in a freestanding midwifery unit, for the subgroup of multiparous women there were no significant differences in the rates. Similarly, for the subgroup of nulliparous women, the rates of episiotomy and third or fourth degree perineal trauma were not significantly different between the two settings. In addition, for nulliparous women the subgroup analysis found that the incidence of the composite neonatal morbidity and mortality outcome was significantly higher in planned home births for all nulliparous low risk women and for the group of nulliparous women without complicating conditions at the onset of labour. Having discussed the results of the subgroup analysis, the group concluded that for multiparous women there was little difference in clinical outcomes between the two settings and therefore that it would come down to the choice of the individual woman. However, they agreed that for nulliparous women, this analysis provided evidence – for this comparison – to support the recommendation that they should be advised to plan to give birth in a freestanding midwifery unit but not at home.

3.2.4. Home compared with alongside midwifery unit

3.2.4.1. Description of included studies

One study was included in this review (Birthplace in England Collaborative Group, 2011).

The included study was a prospective cohort study from England which evaluated outcomes for women intending to give birth at home compared with women intending to give birth in an alongside midwifery unit (Birthplace in England Collaborative Group, 2011).

A summary of points to note about the study population can be found in table 8 below, and further details about the selection of the study groups are reported in the evidence tables (appendix I).

Table 8. Summary of included studies for planned birth at home compared with planned birth in an alongside midwifery unit.

Table 8

Summary of included studies for planned birth at home compared with planned birth in an alongside midwifery unit.

3.2.4.2. Evidence profile

All risk ratios were calculated as standard using RevMan. For measures of perinatal/neonatal mortality and morbidity, due to the low incidence absolute effects have been reported per 1,000,000.

Table 9. Summary GRADE profile for comparison of planned birth at home with planned birth in an alongside midwifery unit for all women.

Table 9

Summary GRADE profile for comparison of planned birth at home with planned birth in an alongside midwifery unit for all women.

3.2.4.3. Evidence statements

One study (n=33,550) reported this comparison. There was evidence that women planning birth at home had a lower risk of instrumental vaginal birth and a lower risk of caesarean section, and therefore a higher rate of spontaneous vaginal birth, than women planning birth in an alongside midwifery unit. Women planning birth at home also had lower rates of epidural use, blood transfusion, episiotomy and third or fourth degree perineal tears.

In terms of neonatal outcomes, the study found evidence that there was no difference in the risk of a composite perinatal mortality and morbidity outcome between babies born to women planning birth at home and women planning birth in an alongside midwifery unit. However, when sub-group analysis by parity was undertaken, there was no difference seen between the two groups for babies born to multiparous women but there was a higher incidence of composite adverse neonatal outcome seen in babies born to nulliparous women planning birth at home compared with nulliparous women planning birth in an alongside midwifery unit. There was no evidence of a difference between the groups in terms of rates of admission to NICU and risks of stillbirth or early neonatal death, but stillbirth and early neonatal death formed part of the composite outcome and the study was not powered to detect a difference in the individual components. The study reported a higher risk of a clinical diagnosis of neonatal encephalopathy (another component of the composite outcome) among babies born to women planning birth at home, but no difference in the rates of babies with the signs of neonatal encephalopathy. The evidence across all outcomes was of low and very low quality.

Transfer rates were similar in both settings but transfers were about three times more common in nulliparous women than in multiparous women. Over one fifth of the multiparous transfers took place after birth.

3.2.4.4. Evidence to recommendations

3.2.4.4.1. Relative value placed on the outcomes considered

The guideline development group members agreed that it was vital to consider both the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in women's decision-making process. For the baby, they felt that it was important to establish whether there was any risk associated with planning birth at home, and therefore the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. Similarly, the group wanted to ascertain whether there were differences in morbidity for the woman following planned birth at home compared with planned birth in an alongside midwifery unit, given that in the case of an unforeseen emergency, such as a postpartum haemorrhage, transfer from an alongside midwifery unit into an obstetric unit would be likely to be more expedient than transfer from home. The rates of intervention, such as caesarean section and instrumental vaginal birth, were also considered priorities, as they were felt to be important to women and were also associated with morbidity, such as postpartum haemorrhage. The group did not feel that use of epidural anaesthesia was a particularly helpful outcome because it is a matter of personal choice for the woman.

The guideline development group also felt that the rates of transfer were important to consider and would be an important consideration for women planning where to give birth.

3.2.4.4.2. Consideration of clinical benefits and harms

The evidence showed that rates of intervention were lower after planned birth at home when compared with planned birth in an alongside midwifery unit. This was consistent across the outcomes of instrumental vaginal birth, caesarean section, episiotomy and blood transfusion. In addition, it was demonstrated that the incidence of third or fourth degree perineal tears was lower after planned birth at home, although no evidence was available for less severe types of perineal trauma.

The guideline development group then considered the neonatal outcomes reported. They discussed the fact that no significant difference was identified in stillbirth or early neonatal death across the two settings, but noted that the study had not been powered to detect a difference in these rare outcomes. They noted that the Birthplace (2011) study was powered to detect a difference in its composite outcome, but that this had required pooling components with very different levels of severity, from fractured clavicle to mortality. While the group felt that this was a limitation of the data, they also conceded that conducting a study with the power to detect differences in the very severe (and hence rare) mortality outcomes would not be feasible. Also fractured clavicle accounted for less than 3% of events and the serious perinatal outcomes – mortality, neonatal encephalopathy and meconium aspiration – constituted just under 90% of the primary outcome events. No significant difference was identified between the two settings for the composite outcome (across all women) and for rates of admission to NICU, but it was found that clinical diagnosis of neonatal encephalopathy was more common in babies born after a planned home birth. The group felt that this outcome was quite difficult to interpret because it was not split by the grade of encephalopathy. Some of the babies would have had symptoms that resolved with no long term effects, whereas others might have had serious morbidity.

The group noted that the rates of transfer were fairly similar in planned home births and planned births in alongside midwifery units, and that rates of transfer were three times higher in nulliparous women. They also noted that over one-fifth of the multiparous transfers took place after birth.

3.2.4.4.3. Consideration of health benefits and resource uses

The guideline development group discussed the relative costs associated with planning birth at home and in an alongside unit. They noted that a planned home birth was associated with lower rates of instrumental vaginal birth, caesarean section and blood transfusion, and therefore that costs might be lower in that respect. Home birth is also free from associated ‘hotel’ service costs. However, they noted that transfer from home has an associated cost in terms of the ambulance whereas transfer from an alongside midwifery unit does not require a vehicle. In addition, in an alongside unit one-to-one care can be provided by midwives supported by auxiliary staff, which means that the overall staffing costs associated with providing one-to-one care can be lower than for a home birth, where one midwife is required throughout labour and then two midwives need to be present at birth. Home births also require midwives' time to be spent travelling to and from the birth. On occasion, however, in some services more midwives may be present in an alongside midwifery unit than are required to provide one-to-one care for the women in labour. Similarly, more community midwives may need to be on call to provide care than are required, in order to ensure that staffing levels are adequate for the estimated number of births.

3.2.4.4.4. Quality of evidence

Only one included study was available for this comparison. The study was recent, large and conducted in England, but it was an observational study and there were demographic differences between the two study groups which could have affected birth outcomes. The authors performed adjustments for these demographic differences for other comparisons reported in the study, but not for the comparison of home and alongside midwifery units and therefore the evidence was graded as very low quality.

3.2.4.4.5. Other considerations

The group discussed the subgroup analysis by parity that was reported and noted some cases where outcomes were slightly different to the overall analysis. For multiparous women, the main difference was that planning birth at home was no longer associated with a significantly lower rate of blood transfusion. For nulliparous women, there was no significant difference between the two settings in terms of caesarean section rate, incidence of third or fourth degree perineal tears and rate of blood transfusion, and so some of the benefits of a planned home birth that were demonstrated in the overall analysis were not present for nulliparous women. In addition, for nulliparous women a planned home birth was associated with significantly higher rates of the composite neonatal morbidity and mortality outcome. The guideline development group felt that, with reference to this comparison, the evidence suggested that nulliparous women should be recommended to plan birth in an alongside midwifery unit.

3.2.5. Home compared with obstetric unit

3.2.5.1. Description of included studies

Fifteen studies (reported in 16 papers) were included in this review (Ackermann-Liebrich et al., 1996; Birthplace in England Collaborative Group, 2011; Davis et al., 2011 and 2012; de Jonge et al., 2009; de Jonge et al., 2013; Dowswell et al., 1996; Hutton et al., 2009; Janssen et al., 2002; Janssen et al., 2009; Lindgren et al., 2008; Nove et al., 2012; Pang et al., 2002; van der Kooy et al., 2011; Woodcock et al., 1994; Blix et al., 2012).

One of the studies is a pilot randomised controlled trial conducted in England (Dowswell et al., 1996). Three of the included studies are prospective cohort studies; these were conducted in England (Birthplace in England Collaborative Group, 2011), Switzerland (Ackermann-Liebrich et al., 1996) and Canada (Janssen et al., 2002). The remaining 11 studies are retrospective cohorts carried out in 8 different countries: England (Nove et al., 2012), The Netherlands (de Jonge et al., 2009 and 2013; van der Kooy et al., 2011), Sweden (Lindgren et al., 2008), USA (Pang et al., 2002), Canada (Hutton et al., 2009; Janssen et al., 2009), Australia (Woodcock et al., 1994), New Zealand (Davis et al., 2011 and 2012) and Norway (Blix et al., 2012).

All of the studies compared planned birth at home with planned birth at an obstetric unit and analysed data on an intention-to-treat basis, so that women were analysed by their planned place of birth even if they were transferred. Three of the included studies evaluated outcomes by booked place of birth during the antenatal period (Ackermann-Liebrich et al., 1996; Dowswell et al., 1996; Woodcock et al., 1994). One study (Birthplace, 2011) analysed outcomes by the intended place of birth and the start of care in labour. In the remaining studies, outcomes were analysed by the intended place of birth at the onset of labour.

The included studies aimed to restrict their populations to low risk women, but a proportion of women had complications which resulted in them being higher risk or outside the scope of the guideline. There are also systematic differences between the characteristics of the study groups in many of the studies, skewing the results in a certain direction, as the majority of included studies are cohort studies and women planning a home birth are a self-selected group of women. A summary of points to note about the study populations can be found in table 10 below. Further details about the selection of the study groups are reported in the evidence tables (appendix I).

Table 10. Summary of included studies for planned birth at home compared with planned birth in an obstetric unit for all women.

Table 10

Summary of included studies for planned birth at home compared with planned birth in an obstetric unit for all women.

3.2.5.2. Evidence profile

All risk ratios were calculated as standard using RevMan; however, where the authors have reported adjusted measures of effect, these have also been reported in the table. For measures of perinatal/neonatal mortality and morbidity, due to the low incidence absolute effects have been reported per 1,000,000.

Table 11. Summary GRADE profile for comparison of planned birth at home with planned birth in an obstetric unit.

Table 11

Summary GRADE profile for comparison of planned birth at home with planned birth in an obstetric unit.

3.2.5.3. Evidence statements

No incidences of maternal mortality in any setting were reported in the five studies (n=75,578) that reported this outcome. There was consistent evidence that women planning birth at home had a lower risk of caesarean section (n=112,837) and instrumental vaginal birth (n=98,158), and hence higher rates of spontaneous vaginal birth (n=100,184), than women planning birth in an obstetric unit. There was also consistent evidence that women planning birth in an obstetric unit had higher rates of epidural use (n=82,850).

The evidence around blood loss was mixed. Evidence from 4 studies (n=44,307) suggested that the risk of any postpartum haemorrhage was reduced in women planning birth at home, but conversely 2 other studies (n=19,734) reported an increased risk in women planning birth at home. In terms of major postpartum haemorrhage, 3 studies (n=434,008) reported that the risk was reduced in women planning a home birth whereas 2 other studies (n=15,472) did not find a difference between the 2 groups. Two large studies (n=183,018) found that the blood transfusion rate was lower for women planning birth at home, however 2 other studies (n=15,158) reported no difference between groups for this outcome. There was consistent evidence from 6 studies (n=77,542) that the rates of episiotomy were higher among women planning birth in an obstetric unit, and consistent evidence from 2 studies (n=2590) that the chance of an intact perineum was higher among women planning a home birth. In terms of vaginal/perineal tears, most studies (n=53,198) found that the risk was higher among women planning birth in obstetric unit. One study (n=414) found the converse and another study (n=11) did not find a difference, but the latter study had fewer than 10 women in each arm. For third or fourth degree tears, 3 studies (n=38,574) reported an increase in risk among women planning birth in an obstetric unit but 3 studies (n=42,518) reported no difference between the 2 groups.

No studies found evidence of a difference in the risk of stillbirth or perinatal death between babies of women planning birth at home and women planning birth in an obstetric unit. Five studies (n=612,134) also did not report a difference in the risk of neonatal mortality. One study did report an increased risk of neonatal mortality among babies born to women who planned birth at home, but the very serious risk of bias undermined confidence in this finding. Given the rare nature of perinatal mortality outcomes, none of the studies were powered to detect a difference in these outcomes. Similarly, 1 study (n=36,176) found no evidence of a difference in the risk of neonatal encephalopathy, but this formed a component of a composite outcome and therefore the study was not powered to detect a difference.

There was inconsistent evidence around the rate of admission to NICU, but the majority of the evidence suggested that there was no difference in risk, with only 1 study (n=3904) suggesting that the risk was reduced among babies born to women planning a home birth.

One study (n=36,104) reported a composite perinatal mortality and morbidity outcome, for which it was powered. There was evidence of no difference between the 2 groups when all low risk women were considered, but when the analysis was restricted to women without complicating conditions identified at the onset of labour, the risk of an adverse perinatal outcome was found to be increased in babies of women who planned a home birth. When sub-group analysis by parity was undertaken it was found that babies born to nulliparous women planning birth at home were more likely to have a composite adverse neonatal outcome than those born to nulliparous women planning birth in an obstetric unit. This was true for all nulliparous women and for those without complicating conditions at the onset of labour. There was no difference seen between groups for babies born to multiparous women, either for all multiparous women or for those without complicating conditions at the onset of labour. The evidence across all outcomes was of low and very low quality.

The transfer rates were fairly consistent, with most reporting rates about 20–25%. One study reported that transfer rates were nearly 4 times more common in nulliparous women.

The quality of evidence for all of these outcomes was predominantly very low, with some studies rated as low for a few outcomes.

3.2.5.4. Evidence to recommendations

3.2.5.4.1. Relative value placed on the outcomes considered

The guideline development group agreed that it was vital to consider the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in a woman's decision-making process. For the baby, they felt that it was important to establish whether there was any risk associated with planning birth at home, and therefore the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. Similarly, the group wanted to ascertain whether there were differences in morbidity for the woman following planned birth at home compared with planned birth in an obstetric unit, given that in the case of an unforeseen emergency, such as a postpartum haemorrhage, women would have to be transferred from home into hospital. The rates of intervention, such as caesarean section and instrumental vaginal birth, were also considered priorities, as they were felt to be important to women and would be associated with morbidity, such as postpartum haemorrhage. The group did not feel that epidural was a particularly helpful outcome, as it is only available in an obstetric unit and is a matter of personal choice for the woman.

The group also felt that the rates of transfer were important to consider and would be an important consideration for women planning where to give birth.

3.2.5.4.2. Consideration of clinical benefits and harms

The guideline development group considered the evidence around mode of birth and agreed that the data demonstrated conclusively that women planning birth at home had lower rates of instrumental vaginal birth and caesarean section, and consequently higher rates of spontaneous vaginal birth, than women planning birth in an obstetric unit.

Similarly, episiotomy was consistently more common after planning birth in an obstetric unit when compared with planning birth at home, but the evidence for other outcomes linked to perineal trauma was less consistent. For third or fourth degree tears, data from 3 studies suggested that they were more common after planned birth at an obstetric unit, while 3 studies (of which one was the Birthplace study) did not find evidence of a difference. Two studies that reported the outcome of intact perineum found that the rate was significantly higher after planned birth at home and therefore the group concluded that the chance of some trauma might be higher after planned birth in an obstetric unit, but noted the inconsistency in the evidence.

When considering the evidence around blood loss, the group noted that the Birthplace study had not reported rates of postpartum haemorrhage. However, they acknowledged the reasoning behind this decision, namely the potential bias in reporting haemorrhage rates in different birth settings, with the researchers preferring to use the more robust surrogate outcome of blood transfusion for clinically important haemorrhage. The evidence around both ‘any postpartum haemorrhage’ and ‘major postpartum haemorrhage’ was inconsistent, with studies reporting effects in both directions. The group also agreed that there was a potential for bias in the reporting of haemorrhages in different settings, particularly with smaller amounts of bleeding, and therefore that differences had to be interpreted with caution. They noted that the Birthplace study (and 2 others) did not find a difference in the rates of blood transfusion for women planning a home birth compared with women planning birth in an obstetric unit.

The group then considered neonatal outcomes and discussed the fact that there were generally no differences found in rates of perinatal death, stillbirth and neonatal death, but noted that the studies were underpowered to detect differences in such rare outcomes. The group noted that Pang et al. (2002) had found a significantly higher rate of neonatal death in babies born to women with a planned home birth, but their reservations about the methodology of the study (discussed in more detail below) meant that they did not place much weight on this finding. The group considered the composite outcome reported by the Birthplace study and noted that achieving sufficient power in this study had required pooling components with very different levels of severity, from fractured clavicle to mortality. While the group noted that this was a limitation, it was felt to be the best available evidence about the relative risks for babies of planning birth in each setting. Also, fractured clavicle accounted for less than 3% of events, while the serious perinatal outcomes of mortality, neonatal encephalopathy and meconium aspiration constituted just under 90% of the primary outcome events. The group noted that for nulliparous women without complicating conditions at the onset of labour, the evidence suggested that there was a higher risk to babies of planning birth at home, but they concluded that the risk was very low across both settings. The evidence suggested that there was no higher risk to babies of multiparous women planning birth at home

The guideline development group discussed the rates of transfer reported in the studies, and concluded that they were generally quite consistent, at around 20–25%. They noted that in the Birthplace study, 45% of nulliparous women planning birth at home were transferred at some point during or after labour, compared to only 12% of multiparous women. The group agreed that it was important that women were given information about the likelihood that they would need to be transferred based on local figures where possible, and what the reasons for this likelihood might be.

3.2.5.4.3. Consideration of health benefits and resource uses

The guideline development group discussed the relative costs implicit in providing care for women planning birth at home and those planning birth in an obstetric unit. When considering women planning a home birth, they noted that these women had lower rates of interventions such as caesarean section, and that this would likely be associated with a cost saving. Also, there would be a reduction in the ‘hotel’ costs, both from women giving birth at home and from reduced hospital-based postnatal stays due to fewer interventions. However, they recognised that an obstetric unit always has to be available for transfer and that many such transfers have associated costs due to use of an ambulance. Given the high rate of transfer in nulliparous women, they agreed that this cost could be substantial. In terms of staffing, the group noted that more midwives would be required in the community if more women planned home births or births in a freestanding unit.

The health economic issues are complex but it is important to remember that this guideline is providing guidance on the care of low risk women in labour and not maternity services as a whole. The financial considerations may differ depending on geographical location and population density, which means that increasing community-based labour care may be cost effective for a large conurbation but not for a rural setting. In addition, the evidence demonstrates a wide variation in the configuration of both freestanding and alongside midwifery units, and this has to be factored into the health economic considerations. A consultant-led unit is always necessary to support maternity services, so reducing the number of caesarean sections or instrumental deliveries may not make a large difference to the overall cost of the service. However, there are health and cost benefits. Furthermore, reducing the number of interventions may mean that women in an obstetric unit may receive better care.

For the number of non-obstetric unit deliveries to increase, midwifery staff need to be redeployed. Moving individuals from one setting to another is not always easy and some midwives prefer to work in an obstetric unit. However, working in a community setting in a team is very attractive to many midwives because they experience relative independence and the work is varied in nature (including home visits, antenatal clinics and classes as well as intrapartum care). The skills required for community practice are different and there are costs associated with re-training and maintaining skills. However, the Birthplace analysis of established services suggested cost savings from low risk multiparous women giving birth away from obstetric units. So although there may be setup costs associated with increasing non-obstetric unit deliveries, this could offer potential savings in the future.

After consideration of all these factors, the guideline development group concluded that a shift from obstetric unit to ‘non-obstetric’, including community-based, deliveries could be achieved through reorganisation of the service. This would involve redeployment of midwives out of obstetric units, with provision of appropriate training and support. The group thought that moving midwives out of obstetric units and therefore increasing the number of non-obstetric unit or home births in this way would not necessarily lead to higher overall running costs.

3.2.5.4.4. Quality of evidence

With the exception of 1 pilot randomised controlled trial of 11 women, all of the included studies were observational studies and, as a result, the evidence was almost universally of very low quality. The guideline development group discussed the fact that women who plan a home birth are often systematically different to women who plan birth in an obstetric unit, and that these differences may be associated with differences in birth outcomes. They agreed that the results therefore had to be interpreted carefully, with this fact taken into consideration.

The group agreed that, although the studies had received the same ‘grade’, the quality of the studies, and their applicability to low-risk women in England and Wales, still varied considerably. The group noted that in some of the studies, adjustments had been made for the difference in characteristics between the study groups or matching had been performed, and felt that these were the more useful studies. In particular, they agreed that the Birthplace study was informative, as it was recent, conducted in England, the largest study reporting most outcomes, and performed adjustments for demographic differences. However, they also noted that although the Birthplace study was restricted to ‘low risk’ women, 19.5% of women planning birth in an obstetric unit had 1 or more complicating conditions identified at the start of care in labour, compared to 5.5% of the women planning a home birth.

The group noted that in some of the studies, there was a particular risk of bias due to the method of selection of the study groups. They discussed the fact that in Pang et al. (2002), large assumptions had to be made to identify the women with a planned home birth and that this could have affected results if unplanned home births had been mistakenly categorised as being planned. Similarly, in Hutton et al. (2009), a proportion of women had their planned place of birth coded as ‘unknown’ and therefore there was a risk that these women could have been misclassified. The group also discussed the fact that in some of the included studies, the authors had used slightly different criteria to select the 2 groups of women, in particular in Lindgren et al. (2008), where pre-term births were included in the planned home birth group only, and Janssen et al. (2009), where women with previous caesarean section were included in the planned home birth group only. They noted that this was generally a very small proportion of women and that there were demographic differences in other studies that used consistent criteria across groups, but their feeling was that there was a greater risk of bias in these studies and therefore that the results should be interpreted with more caution.

3.2.5.4.5. Other considerations

The guideline development group discussed the fact that the availability of home births and the rate of uptake by women varies considerably across England and Wales. They agreed that areas with higher home birth rates and a well-organised service were likely to have a better quality of midwife care for home births, which they expected would result in better outcomes for women and babies. They discussed the fact that most women do not currently choose to give birth at home, for a variety of reasons, but that it was important that women are aware of it as an option and are supported in their decision.

The group considered the subgroup analysis by parity that was reported in the Birthplace study for all outcomes and in various other studies for specific outcomes (Hutton et al., 2009; Janssen et al., 2002; Janssen et al., 2009; Pang et al., 2002). They noted that for most of the outcomes, the analysis for the individual subgroups was consistent with the overall analysis. However, the group did observe some notable differences. Firstly, they noted that for multiparous women, the risk of blood transfusion and third and fourth degree tears was lower in women planning a home birth (the analysis for all women demonstrated no significant difference). They also noted that for the babies of multiparous women, the risk of admission to NICU was lower after a planned home birth and that there was no significant difference in the risk of the composite perinatal morbidity and mortality outcome, even in the subgroup of women without complicating conditions at the onset of labour. Therefore, their conclusion was that for multiparous women and their babies, planning birth at home was as safe as planning birth in an obstetric unit and that the higher rates of intervention in obstetric units suggested that planning birth outside an obstetric unit was preferable. However, when considering outcomes for babies born to nulliparous women, they noted that there was a consistently higher risk of an adverse outcome after a planned home birth, in all low risk women and the subgroup of women without complicating conditions at the onset of labour. Although the risk was low for the babies in both settings, the group felt that the evidence was sufficient to suggest that nulliparous women should not be recommended to plan birth at home while noting that women who choose to plan birth at home should be supported in this choice.

3.2.6. Freestanding midwifery unit compared with alongside midwifery unit

3.2.6.1. Description of included studies

One study was included in this review (Birthplace in England Collaborative Group, 2011).

The included study was a prospective cohort study from England which evaluated outcomes for women intending to give birth in a freestanding midwifery unit compared to women intending to give birth in an alongside midwifery unit.

A summary of points to note about the study population can be found in table 12 below, and further details about the selection of the study groups are reported in the evidence tables (appendix I).

Table 12. Summary of included studies for planned birth in a freestanding midwifery unit compared with planned birth in an alongside midwifery unit.

Table 12

Summary of included studies for planned birth in a freestanding midwifery unit compared with planned birth in an alongside midwifery unit.

3.2.6.2. Evidence statements

Only 1 study (n=27,992) provided comparative data on planned freestanding midwifery unit and alongside midwifery unit births. The evidence suggested that women planning birth in a freestanding midwifery unit had lower rates of instrumental vaginal birth and caesarean section, and therefore higher rates of spontaneous vaginal birth, than women planning birth in an alongside midwifery unit. Women planning birth in a freestanding midwifery unit were also less likely to have an epidural, less likely to have a blood transfusion and less likely to have trauma in the form of an episiotomy or a third or fourth degree perineal tear. There were no reported incidences of maternal mortality in any setting.

In terms of neonatal outcomes, the study found evidence of no difference in a composite adverse neonatal outcome between babies born to women planning birth in a freestanding midwifery unit and babies born to women planning birth in an alongside midwifery unit. A sub-group analysis by parity also found no difference between groups for babies born to either nulliparous or multiparous women. There was also no evidence of a difference in stillbirth, early neonatal death and neonatal encephalopathy, but these outcomes formed part of the composite outcome and the study was not powered to detect a difference in the individual components. There was no evidence of a difference in the risk of admission to NICU between the two groups of babies.

Transfer rates were similar in the two settings. Transfers in nulliparous women were 3 to 4 times more common than in multiparous women. The evidence across all outcomes was of low and very low quality.

3.2.6.3. Evidence to recommendations

3.2.6.3.1. Relative value placed on the outcomes considered

The guideline development group agreed that it was vital to consider both the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in women's decision-making process. For the baby, they felt that it was important to establish whether there was any difference in risk between planning birth in a freestanding midwifery unit and planning birth in an alongside midwifery unit, given that the latter is based in a hospital. Therefore, the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. Similarly, the group wanted to ascertain whether there were differences in morbidity for the woman following planned birth in a freestanding midwifery unit compared with planned birth in an alongside midwifery unit, given that in the case of an unforeseen emergency, such as a postpartum haemorrhage, transfer from an alongside midwifery unit into an obstetric unit would be more likely to be expedient than transfer from a freestanding facility. The rates of intervention, such as caesarean section and instrumental vaginal birth, were also considered priorities, as they were felt to be important to women and also associated with morbidity, such as postpartum haemorrhage. The group did not feel that using epidural anaesthesia was a particularly helpful outcome, as it is a matter of personal choice for the woman and would require a transfer from either setting.

The group also felt that the rates of transfer were important to consider and would be an important consideration for women planning where to give birth. In particular, they noted that, for many women, transfer from a freestanding unit required an ambulance and was likely to be a more stressful and negative experience for the woman than transfer from an alongside unit.

3.2.6.3.2. Consideration of clinical benefits and harms

The guideline development group considered the evidence around outcomes for the woman, and agreed that it universally showed a benefit of planning birth in a freestanding midwifery unit. Women planning birth in a freestanding midwifery unit had higher rates of spontaneous vaginal birth and lower rates of instrumental vaginal birth, caesarean section, blood transfusion, episiotomy and third or fourth degree perineal tears when compared with women planning birth in an alongside midwifery unit. Although it is acknowledged that the data shown in table 13 are not adjusted for parity or other maternal characteristics, the rates of transfer from an alongside midwifery unit were slightly lower (although reasonably similar at 22% compared with 26%) from freestanding midwifery units. Also, women with planned births in alongside midwifery units were typically not transferred to the obstetric unit if the baby required admission to the neonatal unit. In both types of unit, the rates of transfer were 3 to 4 times higher for nulliparous women than for multiparous women, but the group felt that, on balance, the advantages of a freestanding unit outweighed any potential harms and nulliparous women wishing to given birth in freestanding units should be supported in this decision.

Table 13. Summary GRADE profile for comparison of planned birth in a freestanding midwifery unit with planned birth in an alongside midwifery unit for all women.

Table 13

Summary GRADE profile for comparison of planned birth in a freestanding midwifery unit with planned birth in an alongside midwifery unit for all women.

The evidence did not find a significant difference between the two settings in any of the reported perinatal outcomes, although the group noted that the study was only powered to detect a difference in the rate of the composite outcome, not the rare, serious components of stillbirth, neonatal mortality and neonatal encephalopathy. They discussed the fact that this was a limitation of the evidence available, but agreed that conducting a study large enough to detect a difference in the rare outcomes was unlikely to be feasible.

3.2.6.3.3. Consideration of health benefits and resource uses

The guideline development group discussed the relative costs of planning birth in the 2 types of midwifery unit. They agreed that the reduced rate of intervention (such as caesarean section and blood transfusion) among women planning birth in a freestanding unit would reduce the associated costs. However, they also noted that transfer from a freestanding unit required an ambulance and that this would increase the cost when compared to an alongside unit. They also discussed the fact that in areas with small populations, the provision of both types of midwifery unit within 1 region or trust might not be economically viable because the number of places available in midwifery units might exceed the demand. They agreed that, in principle, it was preferable that women have the choice, and felt it would be possible for all areas to provide the option of both types of midwifery-led units, either within 1 region or by working in networks or in collaboration with neighbouring healthcare providers. However, the success of this approach depends on the population density. It may be much easier to achieve in large conurbations than rural areas.

3.2.6.3.4. Quality of evidence

Only 1 included study was available for this comparison. The study was a recent, large prospective study that was conducted in England, but it was an observational study and there were differences between the 2 study groups which could have affected birth outcomes. For example, 6.9% of women in the alongside midwifery unit group had complicating conditions at the start of care in labour when compared with 5.5% of women in the freestanding midwifery unit group. The group also noted that the authors of the study had performed adjustments for demographic differences for other comparisons reported in the study, but not for this one. They agreed that this was a limitation, but were also aware that in fact these 2 groups were quite similar demographically and so any inherent bias caused by these differences was likely to be very small.

3.2.6.3.5. Other considerations

While the Birthplace study findings suggest that intervention rates are lower in planned freestanding midwifery unit births compared with planned alongside midwifery unit births, it has to be acknowledged that the data presented in table 13 are not adjusted for parity or other maternal characteristics and the confidence intervals do not take account of clustering. In the light of this, the guideline development group discussed the subgroup analysis by parity that was reported and agreed that the trends were broadly similar to the results of the overall analysis. The main differences they noted were that when stratified by parity, the difference between freestanding and alongside midwifery units in terms of rates of caesarean section and rates of blood transfusion were no longer statistically significant.

3.2.7. Freestanding midwifery unit compared with obstetric unit

3.2.7.1. Description of included studies

Eight studies were included in this review (Birthplace in England Collaborative Group, 2011; David et al., 1999; Davis et al., 2011; Feldman and Hurst, 1987; Jackson et al., 2003; Overgaard et al., 20112; Scupholme et al., 1986; Stone, 1998).

Five of the included studies are prospective cohort studies, conducted in England (Birthplace in England Collaborative Group, 2011), Denmark (Overgaard et al., 2011) and the USA (Jackson et al., 2003; Scupholme et al., 1986; Stone, 1998). The remaining 3 are retrospective cohort studies, conducted in Germany (David et al., 2011), the USA (Feldman and Hurst, 1987) and New Zealand (Davis et al., 2011).

All of the studies compared planned birth in a freestanding midwifery unit or birth centre with planned birth at an obstetric unit, and analysed data on an intention-to-treat basis, so that women were analysed by their planned place of birth even if they were transferred. One of the included studies evaluated outcomes by booked place of birth during the antenatal period (Jackson et al., 2003), whereas in the remaining studies, outcomes were analysed by the intended place of birth at the onset of labour.

The included studies aimed to restrict their populations to low risk women, but some women had complications which resulted in them being higher risk or outside the scope of the guideline. In addition, there were some systematic differences in the characteristics of women planning birth in a freestanding midwifery unit and women planning birth in an obstetric unit. Three studies aimed to control for potential confounders by performing adjusted analyses (Birthplace in England Collaborative Group, 2011; Davis et al., 2011; Jackson et al., 2003) and a further 3 studies aimed to minimise the differences between the study groups by matching women on socio-demographic characteristics (Scupholme et al., 1986), risk criteria (Feldman and Hurst, 1987) or a combination of different factors (Overgaard et al., 2012). Table 14 shows a summary of points to note about the study populations. Further details about the selection of the study groups are reported in the evidence tables (appendix I).

Table 14. Summary of included studies for planned birth in a freestanding midwifery unit compared with planned birth in an obstetric unit.

Table 14

Summary of included studies for planned birth in a freestanding midwifery unit compared with planned birth in an obstetric unit.

3.2.7.2. Evidence profile

All risk ratios were calculated as standard using RevMan, but where the authors have reported adjusted measures of effect these have also been reported in the table. Some of the included studies only reported percentages without raw event rate data. In order to facilitate analysis, event rate data has been calculated by the technical team where possible (studies where this has been done are designated with a footnote in the table), but in some cases this was not possible due to rounding or unclear reporting of denominators. For measures of perinatal/neonatal mortality and morbidity, due to the low incidence absolute effects have been reported per 1,000,000.

Table 15. Summary GRADE profile for comparison of planned birth in a freestanding midwifery unit and planned birth in an obstetric unit for all women.

Table 15

Summary GRADE profile for comparison of planned birth in a freestanding midwifery unit and planned birth in an obstetric unit for all women.

3.2.7.3. Evidence statements

There was consistent evidence from over 50,000 women that women planning birth in freestanding midwifery units had higher rates of spontaneous vaginal birth and lower rates of instrumental vaginal birth than women planning birth in an obstetric unit. The evidence around caesarean section was slightly more mixed, but most studies reported that rates were higher in women planning birth in an obstetric unit and those that did not find a difference still demonstrated the same trend. There was consistent evidence from 4 studies (n=35,521) that women planning birth in a freestanding midwifery unit had lower rates of epidural use.

The evidence around blood loss was mixed. There was consistent evidence from 2 studies (n=17,888) that there was no difference in the risk of major postpartum haemorrhage between the 2 groups, but another study (n=30,809) reported that the rate of blood transfusion was higher in women planning birth in an obstetric unit.

There was consistent evidence from 6 studies (n=47,163) that rates of episiotomy were higher among women planning birth in an obstetric unit. Similarly, fewer women were found to have an intact perineum or birth canal after planning birth in an obstetric unit (n=18,037). In terms of vaginal or perineal trauma, the evidence was more mixed. There was consistent evidence of no difference in the risk of third or fourth degree trauma (n=49,038), but the evidence around the rates of any trauma was varied, with some studies (n=255) suggesting higher rates in women planning birth in freestanding unit while another large study (n=1678) suggests the converse.

One Danish observational study (n=1678; Overgaard et al., 2012) reported that women in freestanding units were less likely to sustain perineal tears or need a caesarean section than women in obstetric units. Apgar scores were very similar.

Evidence from 2 studies (n=5914) suggested that there was no difference in maternal complications between the 2 groups of women. There was further evidence from another study (n=1678) that the rate of maternal readmission or outpatient visit postpartum was reduced among women planning birth in a freestanding midwifery unit, with a further study demonstrating the same trend.

In terms of neonatal outcomes, 1 study (n=30,750) found evidence of no difference in a composite perinatal mortality and morbidity outcome among babies born to women planning birth in a freestanding midwifery unit compared with those born to women planning birth in an obstetric unit. A sub-group analysis by parity also showed evidence of no difference between groups for either nulliparous or multiparous women. Multiple studies with over 30,000 women did not find evidence of a difference in the rates of stillbirth or neonatal death, but none of the studies were powered to detect a difference in these rare outcomes. Similarly, 1 study (n=30,797) reported no evidence of a difference in the risk of neonatal encephalopathy but it was a component of the composite outcome, and therefore the sample size was insufficient to detect a difference. The evidence around the relative risks of admission to NICU was mixed: 4 studies with over 4000 women did not report a difference in the rates between the 2 groups of babies, but 2 studies with over 30,000 women reported that rates were higher among babies born to women planning birth in an obstetric unit.

Transfer rates varied from about 10–20%. One study (n=64,538) reported transfer rates for nulliparous women were 4 times those in multiparous women. The evidence across all outcomes was of low and very low quality.

3.2.7.4. Evidence to recommendations

3.2.7.4.1. Relative value placed on the outcomes considered

The guideline development group agreed that it was vital to consider both the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in women's decision-making process. For the baby, they felt that it was important to establish whether there was any risk associated with planning birth outside an obstetric unit and therefore the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. Similarly, the group wanted to ascertain whether there were differences in morbidity for the woman following planned birth in a freestanding midwifery unit compared with planned birth in an obstetric unit, given that in the case of an unforeseen emergency, such as a postpartum haemorrhage, women would have to be transferred by ambulance from a freestanding unit into hospital. The rates of intervention, such as caesarean section and instrumental vaginal birth, were also considered priorities, as they were felt to be important to women and also themselves associated with morbidity, such as postpartum haemorrhage. The group did not feel that use of epidural anaesthesia was a particularly helpful outcome, as it is only available in an obstetric unit and is a matter of personal choice for the woman.

The group also felt that the rates of transfer were important to consider and would be an important consideration for women planning where to give birth.

3.2.7.4.2. Consideration of clinical benefits and harms

The guideline development group first considered the outcomes for the woman. The group agreed that the evidence around mode of birth suggested that spontaneous vaginal birth was more likely, and instrumental vaginal birth and caesarean section were less common, for women planning birth in a freestanding midwifery when compared with women planning birth in an obstetric unit. They noted that 2 of the studies (David et al., 1999; Feldman & Hurst, 1987) did not find a significant difference in the rate of caesarean section, but concluded that this was likely to be a result of the small sample sizes.

The group then considered the evidence around blood loss. They noted that Overgaard et al. (2011) had found a significantly lower rate of any postpartum haemorrhage in women who had planned birth in a freestanding unit, but had not found a significant difference in the rate of postpartum haemorrhages of over 1000 ml. Furthermore, they noted that the Birthplace study (2011) had found that blood transfusions were more common for women planning birth in an obstetric unit, and concluded that this was likely to be as a result of the increased rate of operative delivery.

The group noted that evidence around outcomes linked to vaginal and perineal trauma was less consistent. Data from 3 studies suggested that vaginal/perineal tears might be more common after planned birth in a freestanding midwifery unit, but another large study (Overgaard et al., 2011) suggested the opposite. However, given that episiotomy was consistently more common in women planning birth in an obstetric unit, and that the rate of intact perineum was higher in women planning birth in a freestanding midwifery unit (reported in 4 studies, all of which found a significant difference), the group concluded that women planning birth in an obstetric unit were likely to have higher overall rates of vaginal/perineal trauma than women planning birth in a freestanding midwifery unit.

The group then considered the neonatal outcomes reported. They discussed the fact that no significant difference was identified in stillbirth or early neonatal death between the 2 settings, but that none of the studies had been powered to detect a difference in these rare outcomes. They noted that the Birthplace study (2011) was powered to detect a difference in its composite outcome and had not found a significant difference between the two settings. They discussed the fact that the composite outcome comprised outcomes with very different levels of severity (from fractured clavicle to mortality) and agreed that this was a limitation, but agreed that conducting a study with sufficient power to detect the components would probably not be feasible. Also, fractured clavicle accounted for less than 3% of events and the serious perinatal outcomes – mortality, neonatal encephalopathy and meconium aspiration – constituted just under 90% of the primary outcome events. The group noted that while 4 of the studies had not found a difference in the rate of admission to NICU, the 2 largest studies provided evidence that admission to NICU was more common after planned birth in an obstetric unit compared with planned birth in a freestanding midwifery unit. They agreed that this might be partially a result of the difference in the proportion of women with complicating conditions at the onset of labour in the Birthplace study (2011), and could also be linked to the relative ease with which babies could be admitted to NICU from each birth setting. Alternatively, this might be a reflection of iatrogenic harm to newborn babies associated with the increased intrapartum intervention rates associated with giving birth in an obstetric unit or it could be that some obstetric units have a practice of precautionary admission to the NICU.

The group discussed the rates of intrapartum transfer reported in the studies, and noted that they ranged from about 10% to over 20%. They discussed the fact that in the Birthplace study (2011), 36% of nulliparous women planning birth in a freestanding unit were transferred during labour or after birth, compared with only 9% of multiparous women. The group agreed that it was important that women were given information about the likelihood that they would need or choose to be transferred, and what the reasons for this might be. They did not believe, however, that this was a reason not to support nulliparous women in their choice of giving birth in freestanding midwifery units, with the benefits in terms of reduced intervention outweighing any potential risks.

3.2.7.4.3. Consideration of health benefits and resource uses

The guideline development group discussed the relative costs of planning birth in either a freestanding midwifery unit or an obstetric unit and agreed that the former was likely to be more cost effective, given the reduced rate of intervention (such as caesarean section and blood transfusion). However, they noted that transfer from a freestanding unit required an ambulance, which would have an associated cost. The group also discussed the fact that the cost effectiveness of freestanding units would depend quite a lot on the demand and the number of births taking place in the units, because if they were staffed all the time but not used, this would be a waste of resources. They agreed that this would have to be considered carefully when services were being organised. Similarly, they noted that freestanding midwifery units cannot exist in isolation and therefore that some of the cost of maintaining an obstetric unit facility had to be considered in conjunction with the direct costs of the midwifery unit. Despite this, the group concluded that there were likely to be cost savings associated with more women giving birth in a freestanding midwifery unit. The group was also aware of the possibility of some freestanding midwifery units being operated on a model whereby they could be empty when no women were in labour and opened when required, thus minimising running costs and ensuring staffing always matched demand. They felt that the service could be reorganised to move some midwives out of obstetric units into freestanding units to care for low-risk women, and that the lower rates of intervention (that is, caesarean section and instrumental vaginal birth) associated with planning birth in a freestanding midwifery unit would reduce rates overall and therefore doctors' skills in obstetric units would be focussed on caring for women who require this medical expertise because they are at high risk of developing complications and low risk women who required transfer from other settings.

3.2.7.4.4. Quality of evidence

The guideline development group agreed that the quality of the evidence available for this review varied considerably. In particular, the relevance of the evidence to low risk women planning place of birth in England and Wales was affected by the inclusion of some higher risk women in some of the studies and the fact that some studies were conducted in countries such as the USA, where the care of women in labour is not directly comparable to the UK. There are several possible reasons for this, including important differences in midwifery training and practice. The group noted that there were often significant differences in the characteristics of the women planning birth in freestanding midwifery units and those planning births in obstetric units. While the identified differences had been accounted for in several of the studies, they agreed that there was a possibility that other, unknown differences could have affected outcomes. The group felt that the Birthplace study (2011) was generally a good quality observational study because it did perform adjustments for confounders, was conducted more recently and was based in England. They did note that there were more women with complicating conditions identified at the start of care in labour in the obstetric unit group, which might have affected outcomes, but overall the study was felt to be particularly helpful in informing recommendations, given its size and setting. Finally the group noted that there were low event rates for stillbirth and wide 95% confidence intervals as a result.

3.2.7.4.5. Other considerations

The guideline development group discussed the subgroup analysis by parity that was reported in the Birthplace study (2011) and agreed that the trends were broadly similar to the results of the overall analysis. The differences they noted were that for multiparous women, rates of third or fourth degree perineal trauma were significantly lower in women planning birth in a freestanding midwifery unit compared with women planning birth in an obstetric unit, and admissions to NICU were not significantly different between the 2 settings for babies born to multiparous women.

3.2.8. Alongside midwifery unit compared with obstetric unit

3.2.8.1. Description of included studies

Thirteen studies were included in this review (Begley et al., 2011; Bernitz et al., 2011; Birthplace in England Collaborative Group, 2011; Byrne et al., 2000; Campbell et al., 1999; Chapman et al., 1986; Eide et al., 2009; Gaudineau et al., 2013; Hundley et al., 1994; Klein et al., 1984; MacVicar et al., 1993; Waldenstrom et al., 1997; Waldenstrom and Nilsson, 1997).

Nine of the studies reported the results of 8 randomised controlled trials, which were conducted in England (Chapman et al., 1986; MacVicar et al., 1993), Scotland (Hundley et al., 1994), Ireland (Begley et al., 2011), Norway (Bernitz et al., 2011), Sweden (Waldenstrom et al., 1997; Waldenstrom and Nilsson, 1997 [two reports of the same trial]), Canada (Klein et al., 1984) and Australia (Byrne et al., 2000). The remaining 4 studies were prospective observational studies, which were conducted in England (Birthplace in England Collaborative Group, 2011; Campbell et al., 1999), France (Gaudineau et al., 2013) and Norway (Eide et al., 2009).

All of the studies compared planned birth at an alongside midwifery unit with planned birth at an obstetric unit, and analysed data on an intention-to-treat basis, so that women were analysed by their planned place of birth even if they were transferred. Seven of the included studies evaluated outcomes by booked place of birth during the antenatal period (Begley et al., 2011; Byrne et al., 2000; Campbell et al., 1999; Hundley et al., 1994; MacVicar et al., 1993; Waldenstrom et al., 1997; Waldenstrom and Nilsson, 1997), whereas in the remaining studies, outcomes were analysed by the intended place of birth at the onset of labour or at the start of care at the onset of labour.

The included studies restricted their populations to low risk women, but because many of the studies allocated women during the antenatal period, a proportion of women developed complications prior to the onset of labour which resulted in them being higher risk and/or outside of the scope of the guideline (for example women whose labour was induced). Table 16 reports the main issues to note. Full details of the inclusion and exclusion criteria can be found in the evidence tables in appendix I.

Table 16. Summary of included studies for planned birth in an alongside midwifery unit compared with planned birth in an obstetric unit.

Table 16

Summary of included studies for planned birth in an alongside midwifery unit compared with planned birth in an obstetric unit.

3.2.8.2. Evidence profile

A fixed effects model was used for all meta-analyses, except for 1 outcome (intact perineum) where the heterogeneity was high (I2>60%) and so a random effects model was used. All risk ratios were calculated as standard using RevMan. However, where the authors have reported adjusted measures of effect, these have also been reported in the table. For measures of perinatal/neonatal mortality and morbidity, due to the low incidence absolute effects have been reported per 1,000,000.

Table 17. Summary GRADE profile for comparison of planned birth in an alongside midwifery unit with planned birth in an obstetric unit for all women.

Table 17

Summary GRADE profile for comparison of planned birth in an alongside midwifery unit with planned birth in an obstetric unit for all women.

3.2.8.3. Evidence statements

There was consistent evidence from a meta-analysis of randomised controlled trials (n=9208) and a large observational study (n=36,378) that women planning birth in an alongside midwifery unit had higher rates of spontaneous vaginal birth and lower rates of instrumental vaginal birth than women planning birth in an obstetric unit. One smaller observational study (n=1206) also found a higher rate of spontaneous vaginal birth in the alongside midwifery unit group, but no difference in rates of instrumental birth. Two other observational studies (n=1937) did not find a difference in either outcome, but they had smaller sample sizes and demonstrated similar trends. A big observational study (n=36,378) suggested a large reduction in the rate of caesarean section following planned birth in an alongside midwifery unit, and the meta-analysis of randomised controlled trials (n=11,298) demonstrated a trend in that direction, a finding mirrored in a smaller observational study. Three additional observational studies (n=3143) did not find a difference in caesarean section rates between the 2 groups.

There was consistent evidence (n=49,286) that planning birth in an alongside midwifery unit is associated with lower rates of epidural use, but no difference in outcomes of blood loss (measured either using any postpartum haemorrhage [n=10,826], major postpartum haemorrhage [n=1111] or need for a blood transfusion [n=41,318]). Women planning birth in an alongside unit had lower rates of episiotomy, but there was evidence (n=8100) from a meta-analysis of an increased risk of any vaginal/perineal tears (although one observational study [n=1482] did not find a difference in the risk of any tears). No difference was found in the proportion of women with an intact perineum or in the risk of third or fourth degree vaginal/perineal tears. There were no incidences of maternal mortality, although this outcome was only reported in 2 studies (n=38,069).

There was consistent evidence (n=30,796) that there was no difference in the rate of admission to NICU between babies born to women planning birth in an alongside midwifery unit and those born to women planning birth in an obstetric unit. In terms of other neonatal outcomes, 1 large observational study (n=36,075) and 1 smaller observational study found evidence that there was no difference in the rate of a composite perinatal mortality and morbidity outcome among babies born to women planning birth in an alongside midwifery unit and those born to women planning birth in an obstetric unit. This was true for both nulliparous and multiparous women. There was also no evidence of a difference in the risk of stillbirth (n=44,315), neonatal death (n=1828) and neonatal encephalopathy (n=36,156), but none of the studies were powered for these rare outcomes individually.

Transfer rates varied from about 20% to over 50%. Transfer rates were higher in nulliparous women than in multiparous women. The evidence across all outcomes was of low and very low quality.

3.2.9. Evidence to recommendations

3.2.9.1. Relative value placed on the outcomes considered

The guideline development group agreed that it was vital to consider the outcomes for the woman and the outcomes for the baby, as they felt that both would play a part in women's decision-making process. In terms of the woman, the group particularly wanted to ascertain whether the rates of intervention (namely instrumental vaginal birth, caesarean section and episiotomy) and the associated morbidities (for example perineal trauma and postpartum haemorrhage) differed in women planning birth in the 2 settings, because they are both based in hospital. For the baby, they felt that it was important to establish whether there was any risk associated with planning birth outside an obstetric unit, and therefore the outcomes relating to neonatal mortality and morbidity (including the risks of admission to NICU) were considered priorities for decision-making. The group did not feel that epidural was a particularly helpful outcome, as it is only available in an obstetric unit and is a matter of personal choice for the woman.

The group also felt that the rates of transfer were important and would be a significant consideration for women planning where to give birth.

3.2.9.2. Consideration of clinical benefits and harms

The guidelines development group noted that for many of the outcomes, the evidence showed that there was no significant difference between planning birth in an alongside midwifery unit and planning birth in an obstetric unit. In particular, the group noted that no statistically significant differences were identified for the neonatal outcomes reported, but they did discuss the fact that the studies were not powered to detect differences in individual rare outcomes (stillbirth, neonatal death and neonatal encephalopathy). They noted that the Birthplace study (2011) was powered to detect a difference in its composite outcome, but that this had required pooling components with very different levels of severity, from fractured clavicle to mortality. While the group felt that this was a limitation of the data, they also conceded that conducting a study with the power to detect differences in the very severe (and hence rare) outcomes would not be feasible. Also, fractured clavicle accounted for less than 3% of events and the serious perinatal outcomes – mortality, neonatal encephalopathy and meconium aspiration – constituted just under 90% of the primary outcome events. Given this evidence, and the fact that there was no significant difference in the rates of admission to NICU, they concluded that planning birth in an alongside midwifery unit was as safe for the baby as planning birth in an obstetric unit.

The group then considered the outcomes around mode of birth. Overall, the evidence showed that the rates of spontaneous vaginal birth were higher, and the rates of instrumental vaginal birth and caesarean section were lower, after planning birth in alongside midwifery unit when compared with planning birth in an obstetric unit. It was felt there were several reasons why this might be the case: for example, women who wanted regional analgesia with their higher likelihood of instrumental birth would be delivered in an obstetric unit, or, if a woman was not in an obstetric unit with easy access to instrumental birth, more efforts may be made to achieve a spontaneous vaginal birth. They discussed the fact that 2 of the observational studies did not find statistically significant differences between modes of birth, but agreed that this was likely to be a function of the small size of the studies.

In terms of maternal morbidity, the group noted that there was no evidence of a difference in blood loss, in terms of either postpartum haemorrhage or blood transfusion, between women who planned birth in the 2 settings. The group then discussed the outcomes linked to vaginal and perineal trauma. They noted that rates of episiotomy were lower after planned birth in an alongside midwifery unit but that, conversely, the rates of vaginal/perineal tears were higher (although there was no significant difference found in the rates of third or fourth degree trauma). Therefore, they concluded that, on balance, there was unlikely to be a substantial difference between the two settings. This was supported by the meta-analysis of five randomised controlled trials, which showed no significant difference in the proportion of women with an intact perineum.

The group considered the rates of transfer reported in the studies. They discussed the fact that the proportion of women transferred was generally higher in the studies which evaluated outcomes by where women had booked to give birth (rather than where they intended to give birth at the onset of labour), because many of the transfers of care occurred antenatally. They also noted that the transfer rate in nulliparous women was consistently higher than the transfer rate in multiparous women. Apart from parity, the group was aware that maternal age was another confounder influencing transfer rates (Rowe et al., 2012). The group agreed that it was important that women were given information about the likelihood that they would need or choose to be transferred, and what the reasons for this might be. They were also aware of the impact that ambulance transfer in labour can have on the woman's experience and satisfaction (Rowe et al., 2012).

3.2.9.3. Consideration of health benefits and resource uses

The guideline development group felt that planning birth in an alongside midwifery unit was likely to be more cost effective than planning birth in an obstetric unit, given that the rates of intervention were lower and that there was no significant difference identified in neonatal outcomes. However, they also noted that alongside midwifery units cannot exist in isolation and therefore that some of the cost of maintaining an obstetric unit facility had to be considered in conjunction with the direct costs of the alongside unit. Despite this, the group concluded that there were likely to be cost savings associated with more women giving birth in an alongside midwifery unit. They felt that the service could be reorganised to move some midwives out of obstetric units into alongside units to care for low risk women, and that the lower rates of intervention (that is, caesarean section and instrumental vaginal birth) associated with planning birth in an alongside midwifery unit would reduce rates overall and therefore reduce costs associated with these interventions, including reduced length of postnatal stay. The care in obstetric units could be focussed on high risk women and low risk women who required transfer from other settings, ensuring that the appropriate expertise is in the right place for each woman, although it has to be acknowledged that this would have staffing and other resource implications.

3.2.9.4. Quality of evidence

The evidence available for this comparison consisted of both randomised controlled trials and prospective observational studies, and ranged from moderate to very low quality. Unlike the evidence from observational studies, the evidence from randomised trials was meta analysed. Although the evidence from randomised controlled trials is objectively considered more robust, the group noted that many of the trials had methodological issues, were conducted in the 1990s or were conducted outside the UK. In addition, they generally included a proportion of women who were not considered low risk at the start of care in labour. They agreed that these issues somewhat undermined the usefulness of the results when considering low risk women currently planning their place of birth in England and Wales. However, they also noted that in the observational studies, there were significant differences in the characteristics of the study groups. While those that were identified had been accounted for in 2 of the studies, they agreed that there was a possibility that other, unknown differences could have affected outcomes.

The group acknowledged that some of their conclusions were based on the Birthplace (2011) data alone and in some cases these were derived from re-analysis of published Birthplace data in which it has not been possible to take full account of possible confounding and relevant design effects, such as clustering. Nonetheless, the group felt that the Birthplace study was generally a good quality observational study, because it is a very large study, it did perform adjustments for confounders, it was conducted more recently and it was based in England. They noted that there were more women with complicating conditions identified at the onset of labour in the obstetric unit group, which might have affected outcomes, but this was taken into consideration by the authors in the analysis and overall the study was felt to be particularly helpful in informing recommendations, given its size and setting.

The group was disappointed to see that no evidence was available for long-term outcomes relating to place of birth. This lack of data not only impacts on the ability to make informed clinical decisions but also severely limits the usefulness of any health economic evaluation for this topic.

3.2.9.5. Other considerations

The guideline development group noted that the definition of an alongside midwifery unit varied considerably in the included studies and therefore that, in some cases, the setting might not be comparable to alongside units in England and Wales. For example, in some of the studies it was reported that an obstetrician could be called to the unit or that equipment was ‘hidden’ rather than absent. And in some cases the unit was in close proximity and in others, although it was in the same building, it was not adjacent. They agreed that this was not optimal, but they noted that there is variation even within England and Wales and agreed that obstetricians were occasionally called to units instead of women being formally transferred. As a result, it was not possible to define the optimum alongside midwifery unit model to improve outcomes, cost effectiveness and women's experiences. Nevertheless, the group concluded that although the variation in units was a consideration, the results of the review were still valid and applicable to alongside midwifery units in England and Wales. Furthermore, the group questioned whether ‘midwifery-led care’ in an obstetric unit was equivalent in all respects to ‘midwifery care’ in an alongside unit.

The group discussed the subgroup analysis by parity that was reported in the Birthplace study (2011) and two of the randomised controlled trials. They noted that, in general, the results were consistent with those of the overall analysis. The Birthplace study (2011) showed that both nulliparous and multiparous women planning birth in an alongside midwifery unit had higher rates of spontaneous vaginal birth and lower rates of caesarean section, ventouse births and episiotomies than women planning birth in an obstetric unit, with no significant difference shown in neonatal outcomes. The group was disappointed that it was not possible to do a subgroup analysis for the outcome of ‘any vaginal/perineal tears’, although they noted that again no significant difference was found in the incidence of third or fourth degree trauma. The group discussed the fact that the meta-analysis of randomised controlled trials did not show a significant difference in instrumental vaginal birth or caesarean section rate in either nulliparous or multiparous women, but agreed that this was likely to be due to a loss of statistical power.

3.2.10. Health economics profile

Only one economic evaluation of birth place was identified for the UK (Schroeder at al., 2012). A summary of this evaluation is given here and with a full review in appendix A. The aim of the primary analysis was to estimate the cost effectiveness of births planned in different settings (home, freestanding midwifery units, alongside midwifery units and obstetric units) for women and babies at ‘low risk’ of complications prior to the onset of labour.

The study population was all women in the Birthplace prospective cohort study (2011) at ‘low risk’ of complications prior to the onset of labour where the primary outcome and potential confounders were not missing.

The time horizon of the analysis was the duration of follow-up of the Birthplace prospective cohort study (2011). Women were identified at the start of their care in labour and follow-up was complete when the intrapartum and related postnatal care for both mother and baby ended. This time horizon included higher level postnatal or neonatal care but did not include the life-long health effects due to morbidities associated with labour and birth.

The results were presented for all women considered at ‘low risk’, for nulliparous and multiparous women and for women without complicating conditions at the start of care in labour. The cost effectiveness of births planned at home, in freestanding midwifery units and in alongside midwifery units compared with births planned in obstetric units was presented for these separate outcomes:

  • incremental cost per adverse perinatal outcome avoided
  • incremental cost per maternal morbidity avoided
  • incremental cost per additional ‘normal birth’.

When compared with women planning to give birth in an obstetric unit, women planning a birth at home were more likely to

  • be white and have a fluent understanding of English
  • be older (28% were 35 years and older in the home group compared with 16% in the obstetric unit group)
  • be married or living with a partner
  • be living in a more socioeconomically advantaged area
  • have had one or more previous pregnancies.

Each of these characteristics was associated with being cost saving.

The primary outcome was presented in the cost effectiveness analysis as increased or decreased numbers of adverse perinatal outcomes in each setting compared with the obstetric unit. It is important to note that although there is a difference in this outcome between settings, the numbers are very small. The incidence of adverse perinatal outcomes was low in all settings included in the study.

For nulliparous women, a home birth was likely to be less costly but with more adverse perinatal outcomes than birth in other settings for all women considered low risk and for women without complicating conditions at the start of care. Both freestanding midwifery units and alongside midwifery units were less costly and more effective than obstetric units: fewer adverse perinatal outcomes occurred in midwife-led units, although there is uncertainty surrounding the adverse perinatal outcomes.

Table 18. Incremental costs, clinical effects (adverse perinatal outcome avoided) per woman and cost effectiveness ratios (cost per adverse perinatal outcome avoided) for nulliparous women at low risk of complications by planned place of birth using the obstetric unit (OU) as a reference.

Table 18

Incremental costs, clinical effects (adverse perinatal outcome avoided) per woman and cost effectiveness ratios (cost per adverse perinatal outcome avoided) for nulliparous women at low risk of complications by planned place of birth using the obstetric (more...)

Table 19. Incremental costs, clinical effects (adverse perinatal outcome avoided) per woman and cost effectiveness ratios (cost per adverse perinatal outcome avoided) for multiparous women at low risk of complications by planned place of birth using obstetric unit (OU) as a reference.

Table 19

Incremental costs, clinical effects (adverse perinatal outcome avoided) per woman and cost effectiveness ratios (cost per adverse perinatal outcome avoided) for multiparous women at low risk of complications by planned place of birth using obstetric unit (more...)

In terms of incremental cost per maternal morbidity avoided and cost per normal birth, planned place of birth in all non-obstetric unit settings generated incremental cost savings and improved maternal outcomes.

Table 20. Incremental costs, clinical effects (maternal morbidity avoided) per woman and cost effectiveness ratios (cost per maternal morbidity avoided) for all women at low risk of complications by planned place of birth using obstetric unit (OU) as a reference.

Table 20

Incremental costs, clinical effects (maternal morbidity avoided) per woman and cost effectiveness ratios (cost per maternal morbidity avoided) for all women at low risk of complications by planned place of birth using obstetric unit (OU) as a reference. (more...)

Table 21. Incremental costs, clinical effects (normal births) per woman and cost effectiveness ratios (cost per normal birth) for all women at low risk of complications by planned place of birth using OU as a reference.

Table 21

Incremental costs, clinical effects (normal births) per woman and cost effectiveness ratios (cost per normal birth) for all women at low risk of complications by planned place of birth using OU as a reference.

The authors concluded that for multiparous women at low risk of complications, planned birth at home was the most cost-effective option. For nulliparous low risk women, planned birth at home was reported to be the most cost-effective option, although they noted that home births were associated with an increase in adverse perinatal outcomes.

No long-term costs or benefits were included in the analysis. The results were not presented as an incremental cost per quality adjusted life year (QALY). Although the results were not presented in incremental costs per QALY, a willingness to pay threshold of £20,000 was used in the published analysis to make statements about the optimal birth setting by cost effectiveness. As this composite outcome included neonatal death and encephalopathy, as well as fractured clavicle, it is difficult to interpret the results. It implies society is willing to pay £20,000 for 1 adverse perinatal event, where the adverse events vary in severity. Nearly 90% of the composite perinatal outcomes were the more serious outcomes of mortality, neonatal encephalopathy and meconium aspiration.

The same threshold has been applied where effectiveness has increased or decreased relative to the obstetric unit. So the results show home births are less expensive and there are more adverse perinatal outcomes compared with births in the obstetric unit for nulliparous women, whereas the midwifery units are both less expensive and have fewer adverse perinatal outcomes than the obstetric unit. However, if a cost effectiveness threshold of £20,000 per perinatal adverse outcome is applied, then the authors have calculated that home births would result in the highest net benefit. This implies that health decrements can be valued in the same way as health benefits. By applying the same economic value of £20,000 per adverse perinatal event, the increase in adverse perinatal events is acceptable because the cost savings and increased health benefits associated with home births outweigh the costs and health decrements resulting from this increase in adverse perinatal events.

Planned place of birth outside the obstetric unit for low risk women in labour at term (37+0–41+6) is likely to be cost saving. The impact on effectiveness in terms of adverse perinatal outcomes is uncertain. The impact on effectiveness in terms of maternal morbidity and normal births attained show births planned in a setting outside the hospital result in better outcomes.

The authors noted that both costs and cost effectiveness reported may change if maternity services are reconfigured. Commissioners should consider the resource use and related cost implications on the maternity service as a whole. Planning changes to a maternity service to maximise cost effectiveness should be done for a specific local area and use local data to inform decision-making.

Finally, the guideline development group acknowledged that both the contractual model implemented in practice and the providers willingness to ‘absorb’ the risk of lower than expected performance have a great bearing on the cost effectiveness of freestanding midwifery units.

3.2.10.1. Key conclusions (whole review)

The guideline development group noted that the evidence showed that giving birth is generally very safe for women and babies across all settings with very low incidences of mortality or severe morbidity reported.

There was consistent evidence that for women at low risk of complications in birth, giving birth in an obstetric unit is associated with a higher incidence of obstetric intervention and its associated morbidity when compared with planning birth in other settings (though some of these differences may be due to maternal choice, such as an epidural, or transfer in labour, for example because of poor progress). Furthermore, there was evidence of an increase in neonatal unit admissions for babies born in obstetric units with no difference in other neonatal outcomes between planning birth in an obstetric unit, an alongside midwifery unit or a freestanding midwifery unit. Cost effectiveness analysis based on all low risk women also indicates that planning to give birth outside an obstetric unit is more cost effective than planning to give birth in an obstetric unit owing to a reduced incidence of adverse perinatal outcomes.

There were important differences seen for neonatal outcomes when sub-group analyses by parity were undertaken. These data were drawn largely from the UK Birthplace study (2011) and thus were felt to be highly relevant. These findings indicated that for babies born to nulliparous women planning birth at home, there was an increased risk of an adverse composite neonatal outcome when compared with planning birth in all other settings, although the absolute risk of a poor outcome remains small. The guideline development group recognised that the use of a composite outcome for perinatal morbidity could provide misleading results if planned place of birth affects different contributing outcomes in different ways. However, they did not think that was likely. Furthermore, they noted that nearly 90% of the cases with this composite perinatal outcome had conditions which were at the severe end of the spectrum (perinatal death, neonatal encephalopathy or meconium aspiration syndrome). There was no difference seen between settings for babies born to multiparous women. The group noted that this would probably impact on the cost effectiveness of each setting. Although the analysis from Schroeder et al. suggested that planning birth at home would be cost effective for both multiparous and nulliparous women, the group felt it likely that the incremental cost per QALY of giving birth at home would probably be reduced for nulliparous women, given the increased likelihood of an adverse neonatal outcome. As a result, while they recognised that offering both home and midwifery-led settings for multiparous women would be cost effective, they felt less certain about the cost effectiveness of home birth for nulliparous women.

For these reasons, coupled with the fact that planning to give birth in an obstetric unit was found to be associated with increased maternal morbidity, the guideline development group felt it appropriate to recommend that nulliparous women be advised to give birth in either a freestanding or alongside midwifery unit and multiparous women be advised to give birth either at home or in a freestanding or alongside midwifery unit. However, alongside this advice, the group felt it imperative that women be given a choice of all birth settings and be given evidence-based information, in an easily accessible format, about the key risks and benefits associated with each birth setting in order to help them decide the most appropriate place for them to plan to give birth. The group developed a ‘decision aid’ as an example of how this information might be presented to women in a format that facilitates discussion between the woman and the midwife and which the woman can use for future reference and as a basis for ongoing discussions (see appendix L). The group believed it important that the information given should be as relevant as possible to each woman and so recommended that local data be provided wherever possible, and that data relating to nulliparous women and multiparous women be provided separately. Because of this, although the conclusions that inform the recommendations are based on a synthesis of findings from all studies reviewed, the data included in the recommendations are drawn largely from Birthplace (2011), with data from Blix 2012 incorporated for some outcomes for home birth and birth in an obstetric unit (see appendix M for details). The group also drafted a research recommendation to investigate the impact on women's choices and decision-making of using an evidence-based decision aid and a less risk-based approach to discussions around place of birth.

When considering the evidence for each birth setting, a number of gaps in the evidence were noted by the guideline development group, which prompted the drafting of research recommendations. In particular, they drafted a research recommendation to investigate the difference in long-term outcomes associated with birth in the different settings and what components of care in midwifery-led settings contribute to lower intervention rates. They were also aware that the new recommendations would lead to a shift towards more women giving birth outside an obstetric unit than before. Research recommendations were made to look at the staff training needs associated with this shift and to develop a commissioning tool to examine the impact this would have on planning service provision.

3.2.11. Recommendations

  1. Explain to both multiparous and nulliparous women who are at low risk of complications that giving birth is generally very safe for both the woman and her baby. [2014]
  2. Explain to both multiparous and nulliparous women that they may choose any birth setting (home, freestanding midwifery unit, alongside midwifery unit or obstetric unit), and support them in their choice of setting wherever they choose to give birth:
    • Advise low-risk multiparous women that planning to give birth at home or in a midwifery-led unit (freestanding or alongside) is particularly suitable for them because the rate of interventions is lower and the outcome for the baby is no different compared with an obstetric unit.
    • Advise low-risk nulliparous women that planning to give birth in a midwifery-led unit (freestanding or alongside) is particularly suitable for them because the rate of interventions is lower and the outcome for the baby is no different compared with an obstetric unit. Explain that if they plan birth at home there is a small increase in the risk of an adverse outcome for the baby. [new 2014]
  3. Using tables 22 and 23, explain to low-risk multiparous women that:
    • planning birth at home or in a freestanding midwifery unit is associated with a higher rate of spontaneous vaginal birth than planning birth in an alongside midwifery unit, and these 3 settings are associated with higher rates of spontaneous vaginal birth than planning birth in an obstetric unit
    • planning birth in an obstetric unit is associated with a higher rate of interventions, such as instrumental vaginal birth, caesarean section and episiotomy, compared with planning birth in other settings
    • there are no differences in outcomes for the baby associated with planning birth in any setting. [new 2014]
  4. Using tables 24 and 25, explain to low-risk nulliparous women that:
    • planning birth at home or in a freestanding midwifery unit is associated with a higher rate of spontaneous vaginal birth than planning birth in an alongside midwifery unit, and these 3 settings are associated with higher rates of spontaneous vaginal birth than planning birth in an obstetric unit
    • planning birth in an obstetric unit is associated with a higher rate of interventions, such as instrumental vaginal birth, caesarean section and episiotomy, compared with planning birth in other settings
    • there are no differences in outcomes for the baby associated with planning birth in an alongside midwifery unit, a freestanding midwifery unit or an obstetric unit
    • planning birth at home is associated with an overall small increase (about 4 more per 1000 births) in the risk of a baby having a serious medical problem compared with planning birth in other settings. [new 2014]
  5. Ensure that all healthcare professionals involved in the care of pregnant women are familiar with the types and frequencies of serious medical problems that can affect babies (see appendix K), in order to be able to provide this information to women if they request it. [new 2014]
  6. Commissioners and providersl should ensure that all 4 birth settings are available to all women (in the local area or in a neighbouring area). [new 2014]
  7. Give the woman the following information, including local statistics, about all local birth settings:
    • Access to midwives, including:
      • the likelihood of being cared for in labour by a familiar midwife
      • the likelihood of receiving one-to-one care throughout labour (not necessarily being cared for by the same midwife for the whole of labour)
    • Access to medical staff (obstetric, anaesthetic and neonatal).
    • Access to pain relief, including birthing pools, Entonox, other drugs and regional analgesia.
    • The likelihood of being transferred to an obstetric unit (if this is not the woman's chosen place of birth), the reasons why this might happen and
    • the time it may take. Refer to table 26 if no local data are available. [new 2014]
  8. If further discussion is wanted by either the midwife or the woman about the choice of planned place of birth, arrange this with a consultant midwife or supervisor of midwives, and/or a consultant obstetrician if there are obstetric issues. [new 2014]
  9. When discussing the woman's choice of place of birth with her, do not disclose personal views or judgements about her choices. [new 2014]
  10. Ensure that all women giving birth have timely access to an obstetric unit if they need transfer of care for medical reasons or because they request regional analgesia. [new 2014]
  11. Commissioners and providersm should ensure that there are:
    • robust protocols in place for transfer of care between settings (see also recommendations 48 to 52).
    • clear local pathways for the continued care of women who are transferred from one setting to another, including:
      • when crossing provider boundaries
      • if the nearest obstetric or neonatal unit is closed to admissions or the local midwifery-led unit is full. [new 2014]
  12. Commissioners and providersn should ensure that there are multidisciplinary clinical governance structures in place to enable the oversight of all birth settings. These structures should include, as a minimum, midwifery (including a supervisor of midwives), obstetric, anaesthetic and neonatal expertise, and adequately supported user representation. [new 2014]
Table 22. Rates of spontaneous vaginal birth, transfer to an obstetric unit and obstetric interventions for each planned place of birth: low-risk multiparous women (sources: Birthplace 2011; Blix et al. 2012).

Table 22

Rates of spontaneous vaginal birth, transfer to an obstetric unit and obstetric interventions for each planned place of birth: low-risk multiparous women (sources: Birthplace 2011; Blix et al. 2012).

Table 23. Outcomes for the baby for each planned place of birth: low-risk multiparous women (source: Birthplace 2011).

Table 23

Outcomes for the baby for each planned place of birth: low-risk multiparous women (source: Birthplace 2011).

Table 24. Rates of spontaneous vaginal birth, transfer to an obstetric unit and obstetric interventions for each planned place of birth: low-risk nulliparous women (sources: Birthplace 2011; Blix et al. 2012).

Table 24

Rates of spontaneous vaginal birth, transfer to an obstetric unit and obstetric interventions for each planned place of birth: low-risk nulliparous women (sources: Birthplace 2011; Blix et al. 2012).

Table 25. Outcomes for the baby for each planned place of birth: low-risk nulliparous women (source: Birthplace 2011).

Table 25

Outcomes for the baby for each planned place of birth: low-risk nulliparous women (source: Birthplace 2011).

Table 26. Primary reasons for transfer to an obstetric unit (source: Birthplace 2011).

Table 26

Primary reasons for transfer to an obstetric unit (source: Birthplace 2011).

3.2.12. Research recommendations

1. How does the provision of accurate, evidence-based information affect women's decision-making processes and choice of place of birth?

Why this is important

A report by Coxon et al. (2013) identifies in detail why women make choices about where give birth and how these choices can be influenced. Influences may include written and verbal information (both online and from midwives and doctors), previous experience, and word-of-mouth advice from friends and family. The GDG concluded from the Birthplace study that giving birth outside an obstetric unit is the optimal choice for low-risk women. This finding should be used to restructure the way in which information is provided, so that it is presented in a more accurate, less risk-based way in order to support women's choices. This change should be evaluated in a quantitative observational study and/or qualitative study that records any changes in women's choice-making about place of birth. Outcomes include understanding why and how women make choices about where to give birth and how this can influence the provision of appropriate and accessible information, a measure of informed decision-making, and fearfulness and absence of fearfulness when choosing place of birth.

2. What are the long term consequences for women and babies of planning birth in different settings?

Why this is important

The long-term consequences of birth experiences and birth outcomes are poorly understood, particularly in relation to place of birth. A large population-based observational study would compare women's experiences and outcomes for them and their babies in different birth settings (with subgroup analysis by mode of birth) in relation to the wellbeing of the women and their children over different periods of time (for example, 2, 5, 10, 15, 20 and 30 years). A secondary analysis could compare different providers where birth philosophies are different. Outcomes would be compared by accessing medical records and through qualitative interviews. Primary outcomes are mode of birth, long-term morbidity in the woman and baby, pain after birth, readmission to hospital, infection, psychological morbidity (for example, postnatal depression, bonding, relationship breakdown with partner, fear of giving birth in future) and breastfeeding rates. Secondary outcomes are impact on attachment between mother and child, obesity in children, autoimmune disease, chronic illness, educational achievement and family functioning.

3. What are the resourcing and staff training needs associated with advising women at low risk of developing intrapartum complications to give birth outside an obstetric unit and providing a service that facilitates choice of birth setting that includes birth at home, in a freestanding midwifery unit, and alongside midwifery unit and birth in an obstetric unit. What is the impact on satisfaction for women and midwives?

Population: Maternity services providing intrapartum care. Women at low risk of developing intrapartum complications, midwives and other health care professionals involved in providing intrapartum care.

Intervention: Advising women to give birth outside an obstetric unit and providing a choice of birth at home, in a freestanding midwifery unit, an alongside midwifery unit or an obstetric unit.

Comparison: previous model of service provision and uptake including the previous distribution of births between home, freestanding midwifery unit, alongside midwifery unit and obstetric unit.

Outcomes: Resource use, identified training needs and satisfaction for women and midwives

Study design: economic and service evaluation with a qualitative component

Why this is important

Until recently, most births have taken place in an obstetric unit. The evidence relating to safety of birth outside an obstetric unit either at home or in a midwifery-led unit has resulted in a recommendation that low risk women should give birth outside an obstetric unit. This change should lead to a change in service provision by commissioners in order to meet the altered demand. There is also likely to be a change in working practices for midwives, who are more likely to support out of hospital births, and as a result, births that take place within an obstetric unit, although reduced in number, are likely to be women at higher risk of complications. The costs of these changes should be studied, along with satisfaction of both women and midwives. Midwives' training needs should be identified relating to offering choice of place of birth, as well as skills required to work outside a high-risk obstetric unit.

4. Development and validation of a commissioning tool to enable the calculation of the effect of changes in the configuration of the provision of options for place of birth in an individual health economy, network or District

Population: a Health Economy, District or network providing maternity services for a specified population.

Intervention: a change in the provision of alternate places of birth – e.g. the opening of a freestanding and / or alongside midwifery unit or the development of a home birth team

Comparator: current service configuration/provision

Outcome: The costs of provision of maternity services (antenatal care through to postnatal care) for all women giving birth within that maternity service, taking into account alterations in service provision; number of women choosing each birth setting and actual place of birth including transfer rates; rates of normal births, and intervention rates such as emergency caesarean section; rates of maternal and neonatal morbidity and associated hospital admissions.

Study design: service evaluation based on observational data.

Why this is important

The current model of assessing costs associated with place of birth is based on the alterations in intervention rates (predominantly caesarean section) by increasing the number of births outside an obstetric unit. However, releasing those potential savings may be difficult because of the number of fixed costs associated with the obstetric unit. Developing new midwifery units is seen to add costs if corresponding savings are not made in the obstetric unit. Initial capital costs of opening a freestanding midwifery unit, as well as different staffing requirements; more senior midwives may be needed compared to alongside midwifery units or obstetric units; will have an impact on the cost-effectiveness of introducing a new service. It is also important to take into account all uses of an alternative birth setting in terms of antenatal and postnatal care and consider costs to the maternity service as a whole. Commissioners need an economic model that allows them to understand the effect of changes in service provision and how that can alter or not alter the fixed and revenue costs associated with the provision of maternity care for women at high risk of developing complications when developing the provision of services to support birth outside an obstetric unit.

5. What are the key components in midwifery-led settings that result in lower intervention rates?

Population: Women at low risk of developing intrapartum complications.

Intervention: Intrapartum care planned for a midwifery-led birth setting (home, freestanding midwifery unit or freestanding midwifery unit)

Comparison: Intrapartum care planned for an obstetric unit

Primary outcomes: Pain relief; coping strategies offered and used; use of water; 1 to 1 care; type of recordkeeping used; involvement of partner in birth and postnatal period. Philosophy and culture of organisation and in birth setting.

Secondary outcomes: midwives experiences of job satisfaction within midwifery-led environments. Partners experiences of care given. Women and partners' views of environment.

Study design: Observational study with qualitative component

Why this is important

The lower intervention rates observed in midwifery-led settings has been demonstrated consistently across a number of studies. Thus far, however, it has not been conclusively demonstrated what it is about care delivered in these settings that makes this difference. A number of components of care might play a part including one-to-one midwifery care, continuity of carer, intermittent rather than continuous fetal monitoring or a culture that sees labour and birth as a natural process to be worked with rather than ‘treated’ or ‘managed’. A better understanding of what contributes to this observed difference could help improve outcomes and women's experience of care still further, hopefully across all settings.

3.3. Women's experience of planned birth in different settings

3.3.1. Review question

What are women's experiences associated with planning birth in each of the following settings:

  • Home (domicillary)
  • Freestanding midwifery units
  • Alongside midwifery units
  • Obstetric unit/hospital-based maternity unit

For further details on the evidence review protocol, please see appendix E.

3.3.2. Birth planned in an alongside midwifery unit compared with birth planned in an obstetric unit

3.3.2.1. Description of included studies

Seven studies were included in this review (Hundey et al., 2007; Waldenstrom and Nilsson, 1993; Waldenstrom and Nilsson, 1994, Byrne et al., 2000; Littlefield and Adams, 1987; Shaw, 1985; Tingstig et al., 2012). Four of these studies were randomised controlled trials (Hundey et al., 2007; Waldenstrom and Nilsson, 1993; Waldenstrom and Nilsson, 1994, Byrne et al., 2000), 1 was an observational study (Littlefield and Adams, 1987) and 2 were qualitative studies (Shaw, 1985; Tingstig et al., 2012)

One randomised controlled trial conducted in the UK compared birth planned in a ‘midwife-managed unit’ (from this point onwards referred to as ‘alongside midwifery unit’) with birth planned in a ‘consultant-led labour ward’ (from this point onwards referred to as ‘obstetric unit’) (Hundley et al., 1997). The alongside midwifery unit was 5 single rooms in a separate unit located 20 yards from the obstetric unit. The philosophy of care behind the alongside midwifery unit was to provide a safe, ‘homely’ environment where women can retain choice and control in the management of their labours. No further details about the 2 birth settings were reported. The mean age of women in the alongside midwifery group was 28±4.4 years and 56% were nulliparous. The mean age of women in the obstetric unit was 28±4.5 years and 57% were nulliparous. Intrapartum transfer rate was not reported. Outcomes were measured by questionnaires given to women upon discharge from the hospital. Non-responders at 3 weeks were sent the questionnaire again; non-responders at 6 weeks were reminded by telephone. The response rate was 95%.

Two reports of 1 randomised controlled trial conducted in Sweden that compared planned ‘birth centre care’ (from this point onwards referred to as alongside midwifery unit) with planned ‘obstetric care’ (from this point onwards referred to as obstetric unit) (Waldenstrom and Nilsson, 1993; Waldenstrom and Nilsson, 1994). The alongside midwifery unit was located 1 floor below the standard delivery ward but was described as being ‘functionally similar to a freestanding unit with its own staff, facilities and medical guidelines, with an enrolment of 350 women per year’. The alongside midwifery unit offered integrated antenatal, intrapartum and postnatal care in the same premises. The alongside midwifery unit encouraged natural birth and pharmacological pain relief was only available in the case of transfer to the obstetric unit. Fetal monitoring and sonography were not available in the alongside midwifery unit. Access to the alongside midwifery unit was only through participation in the trial. The mean age of women in the alongside midwifery group was 29.9 years (standard deviation not reported) and 57% were nulliparous. The mean age of women in the obstetric unit was 29.7 years (standard deviation not reported) and 53.7% were nulliparous. Intrapartum transfer rate was 17.5% (27.3% of nulliparous women were transferred and 4.5% of multiparous women). Outcomes were measured by questionnaires completed at 2 months after the expected date of birth. The response rate was 93%.

One randomised controlled trial conduced in Australia compared birth planned in a ‘birth centre’ (from this point onwards referred to as alongside midwifery unit) with birth planned in a ‘delivery suite’ (from this point onwards referred to as obstetric unit) (Byrne et al., 2000). The alongside midwifery unit consisted of 2 rooms set up close to the obstetric unit of a maternity teaching hospital. Rooms were ‘homelike’ and all medical equipment was stored behind curtains or in cupboards within easy reach. The mean age of women in the alongside midwifery group was 27.5±5.6 years and 47% were nulliparous. The mean age of women in the obstetric unit was 26.8±4.9 years and 46% were nulliparous. Intrapartum transfer rate was 76%, due to complications or staffing problems. Outcomes were measured by questionnaires completed at 12 hours postpartum. The response rate was 74%.

One observational study conducted in the USA compared birth planned in an ‘alternative birthing unit’ (from this point onwards referred to as alongside midwifery unit) with birth planned in a ‘delivery suite’ (from this point onwards referred to as obstetric unit) (Littlefield and Adams, 1987). The alongside midwifery unit was adjacent to the obstetric unit of a university teaching hospital. The mean age of women in the alongside midwifery unit was 29.6 years (standard deviation not reported) and 29% were nulliparous. The mean age of women in the obstetric unit was 27.1 years (standard deviation not reported) and 77% were nulliparous. Intrapartum transfer rate was 19%. It is important to note that women who transferred out of the alongside midwifery unit were excluded from the analysis. Outcomes were measured 2–3 days postpartum. The overall response rate was 95%.

The findings for these 5 studies are summarised in the GRADE profile in table 27.

Table 27. Summary GRADE profile for comparison of birth planned in an alongside midwifery unit with birth planned in an obstetric unit.

Table 27

Summary GRADE profile for comparison of birth planned in an alongside midwifery unit with birth planned in an obstetric unit.

One observational study, with a qualitative element, that was conducted in Australia aimed to explore the experiences of women who were transferred from an alongside midwifery unit to an obstetric unit (Shaw, 1985). A total of 189 women who had experienced transfer completed questionnaires; 82 of these women (43%) experienced intrapartum transfer. Age and parity were not reported. Questionnaires were completed either at 1 week or 3 months after delivery. Content analysis of questionnaire responses was performed. The findings of this study are summarised in table 28.

Table 28. Qualitative findings for experience of transfer from an alongside unit to an obstetric unit.

Table 28

Qualitative findings for experience of transfer from an alongside unit to an obstetric unit.

One observational study from Sweden was conducted to compare women's satisfaction with ‘modified care’ (in-hospital birth centre) with ‘standard care’ (obstetric unit) (Tingstig et al., 2012). It was a qualitative study, though some frequencies were given. The same medical low risk criteria during pregnancy applied to both groups. For the study, 661 women from the birth centre and 1,097 women from the obstetric unit were invited to take part. The difference in number reflects the researcher's assumption that a larger proportion of the birth centre group would respond than the control group. The 2 groups were each divided into 2 subgroups (nulliparous and multiparous women). Data were elicited 2 months postpartum. The findings are summarised in table 29.

Table 29. Qualitative findings for satisfaction with care in a modified in-hospital birth centre compared with standard care.

Table 29

Qualitative findings for satisfaction with care in a modified in-hospital birth centre compared with standard care.

3.3.2.2. Evidence profile

When appraising the quality of the evidence where outcomes have been measured using a non-standard and simple measure, studies have been downgraded 1 level for risk of (detection) bias. Where it has not been possible to calculate relative and absolute effects, p values have been added to the absolute effect column, when reported in the studies. If a p value is not reported this suggests no significant difference.

Quantitative outcomes are presented under themes of: satisfaction; anxiety and psychological needs; control and decision-making; choice; and support from health professional(s).

3.3.2.3. Evidence statement

3.3.2.3.1. Satisfaction
Overall satisfaction with experience

One randomised controlled trial (n=2844) reported no difference in overall satisfaction with birth experience between women who planned birth in an alongside unit and women who planned birth in an obstetric unit. The evidence for this outcome was of moderate quality.

One qualitative study (n=99) reported that women receiving care in a hospital-based birth centre were more satisfied with overall care than women in an obstetric unit. Being treated as a unique person and a calm atmosphere explained some of the difference in satisfaction levels. The birth centre environment was described as calmer, more personal and less stressful than the obstetric environment.

Comprehensive assessment of satisfaction

One randomised controlled trial (n=1108) reported greater satisfaction in women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of low quality.

Satisfaction with medical supervision and/or treatment

One randomised controlled trial (n=1108) reported greater satisfaction with medical supervision and/or treatment in women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of low quality.

Experience of birth

One randomised controlled trial (n=624) reported no difference in experience of birth between nulliparous women who planned birth in an alongside unit and nulliparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

One randomised controlled trial (n=514) reported no difference in experience of birth between multiparous women who planned birth in an alongside unit and multiparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

Satisfaction with own achievement

One randomised controlled trial (n=624) reported greater satisfaction with own achievement in nulliparous women who planned birth in an alongside unit compared with nulliparous women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of low quality.

One randomised controlled trial (n=484) reported no difference in satisfaction with own achievement in multiparous women who planned birth in an alongside unit compared with multiparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

Happy and satisfied with care

One randomised controlled trial (n=148) reported no difference in the number of women who were happy and satisfied with their care in an alongside unit compared with the number of women who were happy and satisfied with their care in an obstetric unit. The evidence for this outcome was of high quality.

Satisfied with analgesia

One randomised controlled trial (n=148) reported no difference in the number of women who were satisfied with their analgesia in an alongside unit compared with the number of women who were satisfied with their analgesia in an obstetric unit. The evidence for this outcome was of high quality.

Would choose allocated place of delivery for subsequent births

One randomised controlled trial (n=148) reported more women who planned birth in an alongside unit would choose the alongside unit for subsequent births compared with the number of women who planned birth in an obstetric unit who would choose the obstetric unit for subsequent births. This finding was statistically significant. The evidence for this outcome was of high quality.

Satisfaction with delivery environment

One observational study (n=99) reported greater satisfaction with delivery environment in women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Satisfaction with delivery experience

One observational study (n=99) reported greater satisfaction with delivery experience in women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

3.3.2.3.2. Anxiety and psychological needs
Satisfaction with professionals

This considered satisfaction with professional(s) response to women's thoughts and emotions.

One randomised controlled trial (n=1108) reported greater satisfaction with professional(s) response to their thoughts and emotions in women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of low quality.

Anxiety during birth

One randomised controlled trial (n=624) reported no difference in levels of anxiety during birth in nulliparous women who planned birth in an alongside unit and nulliparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

One study (n=514) reported no difference in levels of anxiety during birth in multiparous women who planned birth in an alongside unit and multiparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

3.3.2.3.3. Feelings of control and involvement in decision making
Involvement in labour management decisions

One randomised controlled trial (n=2844) reported no difference in the number of women who were involved in labour management decisions between women who planned birth in an alongside unit and women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

Made own decision about type of pain relief

One randomised controlled trial (n=2381) reported more women who planned birth in an alongside unit made their own decision about the type of pain relief to use compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of moderate quality.

Involvement in the process of birth

One randomised controlled trial (n=624) reported greater involvement in the process of birth in nulliparous women who planned birth in an alongside unit compared with nulliparous women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of low quality.

One randomised controlled trial (n=514) reported no difference in involvement in the process of birth in multiparous women who planned birth in an alongside unit compared with multiparous women who planned birth in an obstetric unit. The evidence for this outcome was of low quality.

Freedom in expression of feelings

One randomised controlled trial (n=624) reported more freedom in expression of feelings in nulliparous women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this finding was of low quality.

One randomised controlled trial (n=514) reported more freedom in expression of feelings in multiparous women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this finding was of low quality.

3.3.2.3.4. Choice
Choice of the way fetal heartbeat monitored

One randomised controlled trial (n=2170) reported more women who planned birth in an obstetric unit had a choice of the way fetal heartbeat was monitored compared with women who planned birth in an alongside unit. This finding was statistically significant. The evidence for this finding was of moderate quality.

Women who wanted to move around and were able to do so most of the time

One randomised controlled trial (n=1317) reported more women who planned birth in an alongside unit who wanted to move around were able to do so compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this finding was of moderate quality.

3.3.2.3.5. Support from health professionals
Support from midwife

One randomised controlled trial (n=624) reported more midwife support was felt to be available by nulliparous women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this finding was of low quality.

One randomised controlled trial (n=514) reported more midwife support was felt to be available by multiparous women who planned birth in an alongside unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this finding was of low quality.

Satisfied with staff

One randomised controlled trial (n=148) reported no difference in the number of women who were satisfied with staff between women who planned birth in an alongside unit and women who planned birth in an obstetric unit. The evidence for this finding was of high quality.

Satisfaction with labour and birth ‘nursing’ care

One qualitative study (n=99) reported no difference in the number of women who were satisfied with labour and birth ‘nursing’ care between women who planned birth in an alongside unit and women who planned birth in an obstetric unit. The evidence for this finding was of very low quality.

Women's views and experiences

One qualitative study (n=82) reported that 45% of women had a positive response to intrapartum transfer from an alongside unit to an obstetric unit, 39% of women had a neutral response and 15% of women had a negative response. They also reported that 7% of women exhibited a great deal of self-criticism in response to intrapartum transfer, 34% some self-criticism and 59% no self-criticism. The evidence for these findings was of very low quality.

3.3.3. Birth planned in a freestanding midwifery unit compared with birth planned in an obstetric unit

3.3.3.1. Description of included studies

Three studies were included in this review (Overgaard et al., 2012; Stone, 1998; Esposito, 1999). Two of these studies were observational studies (Overgaard et al., 2012; Stone, 1998) and 1 was a qualitative study (Esposito, 1999)

One observational study conducted in Denmark compared birth planned in two freestanding midwifery units with birth planned in two ‘supporting’ obstetric units (Overgaard et al., 2012). The freestanding midwifery units were staffed by community midwives working in flexible shifts in a team model and generally providing one-to-one care during labour. Ambulation and use of water and music for pain relief/relaxation were encouraged. In the case of complications, women and babies were transferred to the nearest obstetric unit located 25 to 30 minutes away. In both the freestanding midwifery units and the obstetric units 60% of women were aged under 30 years. Of the women, 23% in the freestanding midwifery units and 24% in the obstetric units were nulliparous. Intrapartum transfer rate was 10%. Outcomes were measured by questionnaires distributed 28 days after birth. The response rate was 86%.

One observational study conducted in the USA compared birth planned in a ‘freestanding birth centre’ (from this point onwards referred to as freestanding midwifery unit) with birth planned in ‘traditional physician care’ (from this point onwards referred to as obstetric unit) (Stone, 1998). Details about the freestanding midwifery unit were not reported. Characteristics of the women included in the study were not clearly reported but authors report women were “generally… educated, married, in their middle to late 20s. Most women had private medical insurance and this was not their first pregnancy”. The intrapartum transfer rate was not reported. The response rate was 77%.

The results of these 2 observational studies are presented in a GRADE table (table 30).

Table 30. Summary GRADE findings for comparison birth planned in a freestanding midwifery unit with birth planned in an obstetric unit.

Table 30

Summary GRADE findings for comparison birth planned in a freestanding midwifery unit with birth planned in an obstetric unit.

One qualitative study conducted in the USA included 29 women either receiving antenatal care or who had given birth at a ‘freestanding birth centre’ located in an inner city neighbourhood (Esposito, 1999). Analysis compared women's birth centre experience with their previous hospital experience. The intrapartum transfer rate was not reported. It was unclear where and when interviews were conducted. The findings of this study are summarised in table 31.

Table 31. Qualitative findings for freestanding midwifery unit experience compared with previous obstetric unit experience.

Table 31

Qualitative findings for freestanding midwifery unit experience compared with previous obstetric unit experience.

3.3.3.2. Evidence profiles

3.3.3.3. Evidence statement

3.3.3.3.1. Satisfaction
Overall birth experience

One study (n=385) reported greater satisfaction with overall birth experience in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Care satisfaction

One study (n=375) reported greater care satisfaction in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Satisfaction with access

One study (n=112) reported greater satisfaction with access in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Satisfaction with ‘nursing care’

One study (n=112) reported no difference in satisfaction with nursing care between women who planned birth in a freestanding midwifery unit and women who planned birth in an obstetric unit. The evidence for this outcome was of very low quality.

Satisfaction with ‘primary care provider’

One study (n=112) reported greater satisfaction with ‘primary care provider’ in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Satisfaction with environment

One study (n=112) reported greater satisfaction with environment in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

3.3.3.3.2. Anxiety and psychological needs
Attention to psychological needs

One study (n=357) reported greater attention to psychological needs in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Feeling of being listened to

One study (n=368) reported a greater feeling of being listened to in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

3.3.3.3.3. Feelings of control and involvement in decision-making
Participation in decision-making

One study (n=356) reported greater participation in decision-making in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Loss of external control over staff actions

One study (n=369) reported less loss of external control over staff actions in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Loss of internal control over labour and own reactions

One study (n=369) reported no difference in loss of internal control over labour and own reactions between women who planned birth in a freestanding midwifery unit and women who planned birth in an obstetric unit. The evidence for this outcome was of very low quality.

Level of information

One study (n=370) reported a greater level of information in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

3.3.3.3.4. Support from health professionals
Support from midwife

One study (n=372) reported greater support from the midwife in women who planned birth in a freestanding midwifery unit compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Midwife present when wanted

One study (n=371) reported that more women who planned birth in a freestanding midwifery unit stated the midwife was present when wanted compared with women who planned birth in an obstetric unit. This finding was statistically significant. The evidence for this outcome was of very low quality.

Women's views and experiences

One study (n=29) found women emphasised the importance of being treated like a person, feeling respected and having their privacy respected, and feeling a connection with care providers during their healthcare experiences in the alongside unit, compared with their previous experiences in an obstetric unit. The evidence for these findings was of very low quality.

3.3.4. Planned home birth compared with birth planned in an obstetric unit

3.3.4.1. Description of included studies

Five studies were included in this review (Janssen et al., 2006; Geerts et al., 2014; Christiaens et al., 2007; Christiaens et al., 2008; Dahlen et al., 2010). Two of these studies were observational studies (Janssen et al., 2006; Geerts et al., 2014) and 3 were qualitative studies (Christiaens et al., 2007; Christiaens et al., 2008; Dahlen et al., 2010).

Two observational studies conducted in Canada and The Netherlands compared planned midwife-attended home births (from this point onwards referred to as home birth) and planned midwife-attended hospital births (from this point onwards referred to as obstetric unit birth) (Janssen et al., 2006 and Geerts et al., 2014). Jansen et al. examined the first 2 years of legislated midwife-attended home births in British Columbia – all women in the province planning a home birth during the study period were required by legislation to enrol in the Home Birth Demonstration Project. The mean age of women in the home birth group was 30.2±5.4 years and 48% were nulliparous. The mean age of women in the obstetric unit group was 31±5.3 years and 56% were nulliparous. The intrapartum transfer rate was 19%. Outcomes were measured by questionnaires, distributed by midwives, and were completed before 6 weeks postpartum. The response rate was 82%. The results of this study are presented in a GRADE table (table 32). Results for the comparison of planned home/actual obstetric unit birth with planned obstetric unit/actual obstetric unit are presented in a second GRADE table (table 33).

Table 32. Summary GRADE profile for comparison planned home birth with planned birth in an obstetric unit.

Table 32

Summary GRADE profile for comparison planned home birth with planned birth in an obstetric unit.

Table 33. Summary GRADE profile for comparison planned home birth/actual obstetric unit birth with planned obstetric unit birth/actual obstetric unit birth.

Table 33

Summary GRADE profile for comparison planned home birth/actual obstetric unit birth with planned obstetric unit birth/actual obstetric unit birth.

Geerts et al. measured feelings of control in women recruited over 1 year in 25 Dutch midwifery practices. Most women were aged between 25 and 35 years (76% in the home birth group and 70% in the hospital birth group). Outcomes were measured by questionnaire using the Likert scale. The results (including transfer frequency) are presented in tables 27 and 29. Hospital obstetric units were predominantly run by midwives, with medical clinicians rarely involved.

Two reports of 1 observational study conducted in Belgium and The Netherlands compared planned home birth and planned hospital birth (from this point onwards referred to as obstetric unit) (Christiaens et al., 2007; Christiaens et al., 2008). The mean age of women in the study was 31.2 years and 54.2% of women in the study were nulliparous (data not reported for home birth and obstetric unit groups separately). Outcomes were measured by questionnaires completed at 2 weeks postpartum. The response rate was estimated to range between 100% and 38% for the home birth group and 68% and 19% for the obstetric unit group. The findings of this study are summarised in table 35.

Table 35. Descriptive findings for home birth and obstetric unit birth.

Table 35

Descriptive findings for home birth and obstetric unit birth.

One qualitative study conducted in Australia included 8 women who had a planned home birth and 8 women who had a planned birth at a tertiary referral public hospital (from this point onwards referred to as obstetric unit) (Dahlen et al., 2010). The study also included 1 woman who gave birth in a private hospital and 2 women who gave birth in a birth centre. The mean age of women in the home birth group was 30 years (standard deviation not reported) while the mean age of women in the obstetric unit group was 25 years (standard deviation not reported). Of the study sample, 89% were nulliparous. One woman who planned a home birth was transferred to hospital (intrapartum transfer). Interviews were conducted in the women's homes between 6 and 26 weeks (mean 15 weeks) after birth. The findings of this study are summarised in table 36.

Table 36. Qualitative findings for home birth and obstetric unit birth.

Table 36

Qualitative findings for home birth and obstetric unit birth.

3.3.4.2. Evidence profiles

Table 34. Summary GRADE profile for comparison: planned home birth (transferred to OU)/planned home birth (not transferred).

Table 34

Summary GRADE profile for comparison: planned home birth (transferred to OU)/planned home birth (not transferred).

3.3.4.3. Evidence statements – planned home birth compared with planned obstetric unit birth

3.3.4.3.1. Satisfaction
Overall satisfaction with childbirth experience

One study (n=658) reported no difference in overall satisfaction with childbirth experience between women who planned a home birth and women who planned birth in an obstetric unit. The evidence was of very low quality.

Total Labour Agentry Scale score

One study (n=658) reported a higher total Labour Agentry Scale score in women who planned a home birth compared with women who planned birth in an obstetric unit (the Labour Agentry Scale is a standardised 29-item scale for measurement of expectancies and experiences of personal control during childbirth). This finding was statistically significant. The evidence was of very low quality.

Women's views and experiences

One qualitative study (n=16) found home birth women felt better equipped to deal with the unknown, which reduced fear. The hospital birth group generally felt less empowered and less well-equipped to deal with the unknown, and consequently talked about fear much more.

Birth setting was not found to be the most important factor in women's birth experiences; instead, it was the care received.

The importance of choice, control and support was talked about by both home birth and hospital birth women. Regardless of birth setting, good communication reduced fear and poor communication and a lack of support increased fear. The evidence was of low quality.

3.3.4.4. Evidence statements – planned home/actual obstetric unit birth compared with planned obstetric unit/actual obstetric unit birth

3.3.4.4.1. Satisfaction
Overall satisfaction with childbirth experience

One study (n=191) reported lower overall satisfaction with childbirth experience in women who planned a home birth and gave birth in an obstetric unit compared with women who planned birth in an obstetric unit and gave birth in an obstetric unit. This finding was statistically significant. The evidence was of very low quality.

3.3.4.4.2. Feelings of control and involvement in decision making
Total Labour Agentry Scale score

One study (n=191) reported no difference in total Labour Agentry Scale score between women who planned a home birth and gave birth in an obstetric unit and women who planned birth in an obstetric unit and gave birth in an obstetric unit. The evidence was of very low quality.

A second study (n=494) reported greater feelings of control in women giving birth at home than women giving birth in hospital, though this was not a significant difference (see evidence table). Women giving birth at home had significantly greater feelings of control if they were not transferred to hospital intrapartum. The evidence was of very low quality.

3.3.5. Birth centre experience compared with previous hospital experience

3.3.5.1. Description of included studies

Two reports of 1 qualitative study conducted in Australia included 17 women who had given birth at 1 of 3 birth centres and had previously given birth in a hospital (Coyle et al., 2001a, Coyle et al., 2001b). It was unclear from the description of the birth centres whether these were alongside or freestanding midwifery units. The women were aged between 22 and 34 years and none were nulliparous. Intrapartum transfer rates were not reported. Interviews were conducted in the women's homes between 2 and 4 months after birth. Analysis compared women's birth centre experience with their previous hospital experience. The findings of this study are summarised in table 37.

Table 37. Qualitative findings for birth centre experience compared with previous hospital birth experience.

Table 37

Qualitative findings for birth centre experience compared with previous hospital birth experience.

3.3.5.2. Evidence profile

3.3.5.3. Evidence statements

3.3.5.3.1. Women's views and experiences

Two reports of one study (n=17) identified different themes for birth centre experience and hospital experience.

3.3.5.3.2. Birth centre experiences: themes

The following key themes were identified for women's experience of birth at a birth centre:

  • Non-interventionist approach: procedures were kept to a minimum.
  • Women as primary decision-makers: midwives provided information to enable women to make informed decisions.
  • Comfortable with carers: continuity of antenatal and intrapartum care facilitated a close relationship between the woman and midwife, which had a positive impact on birth experience.
  • Women ‘being known’: facilitated women's perceptions of ability to be in control of their birth experience.
  • Personalised care and ‘seeing me through’: women felt care was adjusted to them and there was a positive effect from having one midwife carer throughout labour.

The evidence was of low quality.

3.3.5.3.3. Hospital experiences: themes

The following key themes were identified for women's experience of birth in hospital:

  • Interventionist approach: technology was accepted as part of the birth process.
  • Health professional superiority: many women felt that medical experts had a superior attitude.
  • Women as passive participants: women felt discouraged to be involved in decisions, felt a lack of information to facilitate informed decisions.
  • Lack of rapport: some women described unknown carers as a source of anxiety.
  • Women being unknown: carers were unfamiliar with women's birth plans and birth preferences.
  • Systemised care and fragmented labour care: women felt they were ‘just a number’, that the hospital care system was large, inflexible and impersonal, and experienced multiple carers in a short time frame.

The evidence was of low quality.

3.3.6. Experiences of reactions to plans for giving birth at home

One study from Sweden (Sjoblom et al., 2011) elicited women's experiences of reactions to their plans to give birth at home. Reactions came from midwives and other health professionals. Of 1025 women invited to take part, 735 women responded. The women were asked the following question: “If you think anyone has tried to influence you not to give birth at home, would you please describe that experience or the situation you are thinking of?” Of the respondents, 34 women answered that they did not want to talk about it. Answers were analysed using thematic content analysis (Graneim and Lundman, 2008). Table 38 shows each reported category, followed by example quotations.

Table 38. Experience of reactions to plans for home birth.

Table 38

Experience of reactions to plans for home birth.

3.3.6.1. Evidence statements

One large qualitative study (n=735) found that some women were met with negative responses when telling health professionals their plans to give birth at home. They were met with ignorance and prejudice as well as feeling alienated and judged.

3.3.6.2. Health economic profile

No published economic evaluations were identified for this question.

3.3.6.3. Evidence to recommendations

3.3.6.3.1. Relative value placed on the outcomes considered

For this question, the guideline development group wished to consider all outcomes relating to women's experience of labour and birth. The papers reported a range of outcomes under this broad heading but the group did not prioritise any in particular as they were all felt to be relevant. The group had hoped to consider long-term psychological outcomes but none were reported in the studies.

3.3.6.3.2. Consideration of clinical benefits and harms

For the comparison of alongside midwifery units with obstetric units, there were mixed findings. However, generally the results showed that women had a significantly better experience and were more satisfied at 2 months postpartum in the alongside midwifery unit group. For the comparison between freestanding midwifery units and obstetric units, more women felt satisfied, that they had been listened to, that they had been in control of their own decision-making and that they had been supported by a healthcare professional in a freestanding unit. However, the guideline development group recognised that these results were derived from scores on simple scales and that the absolute scores were generally high for both groups. Therefore, the significant differences reported may not be particularly meaningful in terms of women's own experience of care. The group also noted that in studies of this kind not all women will have experienced both types of care setting and this needs to be borne in mind when interpreting the findings of studies.

The group felt that some of the more valuable findings came from the qualitative studies. There were some common themes which came out from a number of the studies about the importance to women of choice, control and being given information. Women spoke positively of experiences where they were provided with sufficient information and were able to make decisions about their care, regardless of the birth setting (although generally these positive experiences were reported in the home and midwifery settings which practiced a non-interventionist model of care). Some studies also showed the importance of providing adequate support to assuage women's concerns and fears and to ensure that the women are sufficiently informed and empowered to make decisions about their care.

The group did not feel that the evidence from this specific review provided sufficient evidence alone to recommend one midwifery setting over another, although they did feel that the evidence showing greater satisfaction with care in midwifery-led settings supported their recommendations regarding place of birth. Furthermore, they agreed that the findings highlighted some extremely important principles of care which all women should be able to expect, regardless of the birth setting, and made recommendations to reflect this.

3.3.6.3.3. Consideration of health benefits and resource uses

As this review question focussed on women's experience of different birth settings, no health economic analysis was performed

3.3.6.3.4. Quality of evidence

The studies ranged in quality from high to very low, although the majority were graded as low or very low. The guideline development group considered the reasons for the studies being downgraded. They recognised that in a number of instances there were methodological flaws. However, overall they felt that there was sufficient evidence to take into account when developing the recommendations.

3.3.6.3.5. Other considerations

The guideline development group recognised that although the evidence demonstrated clear themes in terms of providing supportive care to women, the evidence didn't show how best to ensure that high quality care is put into practice. They acknowledged that providing high quality, individualised care can be challenging, particularly when providing care for large numbers of women and/or when working in a highly pressurised environment. The group discussed this issue and agreed that effective leadership from senior staff can be key in ensuring that high quality care is provided within a maternity service. The group agreed that it was reasonable to recommend this despite the lack of specific evidence since it specifies a level of care that should be provided as standard.

3.3.6.4. Recommendations

13.

For all women giving birth in all birth settings, follow the principles in Patient experience in adult NHS services (NICE clinical guidance 138). [new 2014]

14.

Providers, senior staff and all healthcare professionals should ensure that in all birth settings there is a culture of respect for each woman as an individual undergoing a significant and emotionally intense life experience, so that the woman is in control, is listened to and is cared for with compassion, and that appropriate informed consent is sought. [new 2014]

15.

Senior staff should demonstrate, through their own words and behaviour, appropriate ways of relating to and talking about women and their birth companion(s), and of talking about birth and the choices to be made when giving birth. [new 2014]

3.4. Assessment for choosing place of birth

3.4.1. Review question

What are the risk factors which should be included in assessment to determine the most appropriate place of birth for women during pregnancy and labour?

For further details on the evidence review protocol, please see appendix E.

3.4.2. Description of included studies

No high quality studies were identified that directly addressed this question.

3.4.3. Evidence statement on choosing place of birth

There is no strong evidence on assessment for choosing place of birth and so the guideline development group discussed each condition related to place of birth.

3.4.4. Guideline development group interpretation of the evidence on choosing place of birth

The following criteria have been produced by consensus with the aim of providing consistency of advice for women when considering the relative risk associated with where they wish to give birth.

3.4.5. Recommendations on choosing place of birth

16.

Use tables 39, 40, 41 and 42 as part of an assessment for a woman choosing her planned place of birth:

  • Tables 39 and 40 show medical conditions or situations in which there is increased risk for the woman or baby during or shortly after labour, where care in an obstetric unit would be expected to reduce this risk.
  • The factors listed in tables 41 and 42 are not reasons in themselves for advising birth within an obstetric unit, but indicate that further consideration of birth setting may be required.
  • Discuss these risks and the additional care that can be provided in the obstetric unit with the woman so that she can make an informed choice about planned place of birth. [2007, amended 2014]
Table 39. Medical conditions indicating increased risk suggesting planned birth at an obstetric unit.

Table 39

Medical conditions indicating increased risk suggesting planned birth at an obstetric unit.

Table 40. Other factors indicating increased risk suggesting planned birth at an obstetric unit.

Table 40

Other factors indicating increased risk suggesting planned birth at an obstetric unit.

Table 41. Medical conditions indicating individual assessment when planning place of birth.

Table 41

Medical conditions indicating individual assessment when planning place of birth.

Table 42. Other factors indicating individual assessment when planning place of birth.

Table 42

Other factors indicating individual assessment when planning place of birth.

Footnotes

l

This can also include networks of providers.

m

This can also include networks of providers.

n

This can also include networks of providers.

Copyright © 2014 National Collaborating Centre for Women's and Children's Health.
Bookshelf ID: NBK328267

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (9.4M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...